{"id":"3bdd1bf0-2d77-415b-a3db-ff27005bb519","arxiv_id":"2509.05868","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Charge-state energies and site occupancies of asymmetric four-molecule PTCDA clusters on NaCl/Ag(111) are reproduced by an extended Hubbard model with anisotropic intersite repulsions and hoppings.","lead":"Four negatively charged PTCDA molecules on a salt film form a tiny platform that behaves like an extended Hubbard model, and the authors show that long-range electron repulsion, not the usual on-site repulsion, controls the cluster's behavior. The result suggests molecule-by-molecule assemblies could become tunable simulators for strongly correlated electron systems.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The load-bearing risk is that EFS jumps are interpreted as isolated-cluster addition/removal energies and the equilibrium segment is assigned Ng=2 without an independent charge measurement; a quantitative charge calibration would settle it.","rationale":"The reader's weakest assumption matches the concern I find most load-bearing: the EFS jump-to-energy mapping and the inferred Ng=2. The paper is transparent in stating that Ng is not known from experiment, and the same occupancy data are used to select Ng and to validate the model. A quantitative charge reference would break this circularity. I see no internal inconsistency in the Hamiltonian or in the numerical diagonalization, and the large-U argument that V terms set the transition energies is plausible if the charge-state assignment is right. Because the concern is about underdetermination of the experimental interpretation rather than a demonstrated error, it warrants a conditional rather than a reject verdict; the reader's CONDITIONAL verdict should stand.","tokens_in":11916,"tokens_out":13773,"duration_ms":129721,"concrete_test":"Perform a charge-calibrated EFS measurement on the same surface: use an isolated PTCDA monoanion (known -1e charge) or a Au atom on NaCl(2ML)/Ag(111) as a reference, and extract the absolute contact-potential offset and parabolic frequency-shift curvature over the center of each EFS segment of the diamond cluster. An electrostatic model (tip radius, tip height, dielectric screening of the NaCl bilayer and Ag substrate) calibrated to the reference charge then converts the segment offsets into the total cluster charge. If the equilibrium segment has total charge -2e and the segments reached at +0.15 V and -0.21 V have -3e and -1e, the Ng=2 assignment and the jump interpretation are confirmed. If the equilibrium segment instead has -1e or -3e, the extracted V, VBB, VAA and the \"inter-site terms control occupation\" conclusion must be re-derived under the correct Ng.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The parameter extraction in the Model details section takes the EFS jumps at +0.15 V and -0.21 V as E(Ng+1)-E(Ng) and E(Ng-1)-E(Ng) of the isolated four-site Hamiltonian, and selects Ng=2 by requiring <n_B> greater than <n_A> in the N=2 ground state and <n_A> greater than <n_B> in the N=3 state. The paper states that \"the value of Ng is not known from the experiment\", so the ground-state particle number is inferred from the same jump positions and occupation asymmetries that the model is then said to reproduce. If the jumps are tip-induced charging events (bias-dependent single-electron transfer modulated by the cantilever) rather than ground-state-to-ground-state cluster transitions, or if the true equilibrium segment corresponds to a different Ng, the fitted V, VBB, VAA values and the conclusion that inter-site potentials control the occupation are not determined. The authors also acknowledge in the Discussion that there is no quantitative conversion of surface potential to charge, so the occupancy comparison used to remove parameter degeneracy remains qualitative. This makes the central agreement less independent than it appears and leaves the conclusion dependent on an untested assignment.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports scanning tunneling microscopy, non-contact atomic force microscopy, and electrostatic force spectroscopy measurements on clusters of four PTCDA molecular anions on NaCl/Ag(111), focusing on a C2-symmetric 'diamond' cluster whose EFS spectra show bias-dependent frequency-shift jumps at site-dependent voltages. The authors model the cluster with a four-site extended Hubbard Hamiltonian containing on-site U, nearest-neighbor and diagonal intersite Coulomb repulsions V, VBB, VAA, site-energy offsets εA, εB, and nearest-neighbor plus cross-cluster hoppings t, tAA, tBB. Using exact diagonalization, they determine parameters by requiring the model's electron addition/removal energies to match the EFS jump positions and by imposing site-occupation asymmetries observed in the surface-potential maps. They conclude that, in the large-U limit, the occupation pattern and transition energies are controlled by the non-local site potentials and intersite repulsions rather than by U, and that such molecular-anion clusters are promising platforms for simulating extended Hubbard models.","tokens_in":12215,"tokens_out":4515,"duration_ms":41144,"significance":"If the extracted parameters and their interpretation are independently confirmed, the work would be a valuable experimental realization of an extended Hubbard model in a small molecular cluster, with concrete values for the intersite Coulomb terms (V≈0.315 eV, VBB≈0.281 eV, VAA≈0.247 eV) and a clear demonstration that non-U terms can dominate in the t<<U regime. The manuscript's strengths include a transparently specified Hamiltonian, exact diagonalization rather than approximate methods, an explicit exploration of the tAA/tBB dependence in Fig. 3(d)–(f), and an honest acknowledgment that Ng is not known from experiment and that the surface-potential-to-charge conversion is qualitative. These strengths make the work potentially significant, but the experimental-model link currently contains a circular step and several untested assumptions, so the significance is conditional on additional validation.","major_comments":[{"comment":"The two transition energies E∆N=±1 are used as target values in the constrained optimization: the paper states, 'A constrained optimization algorithm was used to identify the values of V, VAA and VBB that give the correct values of E∆N=±1.' Therefore, the subsequent agreement between the calculated and measured addition/removal energies for this cluster is guaranteed by construction and cannot serve as independent confirmation of the model. Please provide an out-of-sample test or cross-validation—for example, predicting the clover-cluster energy gap using the same intersite potentials, or a quantitative prediction of the N=3 site-resolved occupation from an independent charge calibration—before claiming that the transition energies are well described.","section":"§II B (Model details)"},{"comment":"The paper explicitly states that 'the value of Ng is not known from the experiment,' and Ng=2 is selected because Ng=3 fails the imposed occupation and gap constraints. This makes the ground-state electron number an inference from the same data that are used to fit the model, rather than an independently measured quantity. The extracted values of V, VBB, and VAA, the assignment of the equilibrium EFS segment to N=2, and the central conclusion that intersite potentials control the occupation all rely on this inferred Ng. Please either provide an independent measurement of the cluster charge state (for example, Kelvin-probe or single-electron capacitance measurements) or, failing that, quantify how the fitted parameters and the qualitative conclusions would change if Ng=1 or Ng=3 were assumed, and state this limitation prominently in the Discussion.","section":"§II B and §II C (Ng assignment)"},{"comment":"The EFS frequency-shift jumps are interpreted directly as the isolated-cluster addition and removal energies E(Ng+1)-E(Ng) and E(Ng-1)-E(Ng). However, in EFS on thin insulating films, jump positions can be shifted by tip-induced band bending and cantilever-tunneling coupling (see references [24-26]), and the paper acknowledges this coupling only qualitatively. Please provide a quantitative estimate of tip-induced electrostatic shifts or a control measurement (for example, varying the tip excursion or set-point) to justify treating the jump voltages as ground-state-to-ground-state transition energies of the isolated cluster.","section":"§I (Experimental details) and §II B"},{"comment":"The manuscript asserts that 'the behaviour observed does not depend on U' and that occupation asymmetry is 'independent of U' in the abstract, but the numerical study is performed at a single value of U=1.4 eV. Since this universal-in-U claim is central to the paper's message, please provide a sensitivity analysis over a physically reasonable range of U (for example, 1.2–1.6 eV) showing that the site-occupation asymmetries and the ordering of the addition/removal energies are unchanged, with t<<U maintained.","section":"§II C (Discussion)"}],"minor_comments":[{"comment":"There is a typo in the abstract: 'asymmetric hoping terms' should be 'asymmetric hopping terms'.","section":"Abstract"},{"comment":"The phrase 'Witht << U' is missing a space; it should read 'With t << U'.","section":"Abstract"},{"comment":"The sign convention for εA and εB is confusing: the Hamiltonian contains −εB( n1+n3 )−εA( n2+n4 ), yet the numerical values are given as positive (εA=0.462 eV, εB=0.485 eV). Please clarify whether these are binding energies measured relative to the Fermi level or orbital energies, and how the sign relates to the EFS jump voltages.","section":"§II B, Eq. (1)"},{"comment":"The color scale and the labels 'nB<nA' in panels (b) and (c) of Figure 3 are difficult to parse; the text says the required sign of ⟨nB⟩−⟨nA⟩ is positive in (b) and negative in (c), but the figure caption does not explain the color bar or the overlaid symbols. Please make the sign convention and the plot axes explicit.","section":"Fig. 3"},{"comment":"The statement that the cross-cluster hopping parameters 'allow for delocalization of a single electron across two sites, lowering the overall energy' would be more convincing with a direct energy comparison, showing the energy gain from nonzero tAA/tBB relative to the case tAA=tBB=0.","section":"§II C (Discussion)"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its limitations, which is commendable, but the central validation chain has a circular step: the intersite potentials are fit to the two transition energies that the paper then claims are described by the model, and the ground-state occupation Ng=2 is inferred from the same data. These issues are potentially fixable with additional out-of-sample predictions, a quantitative charge calibration, or sensitivity analyses showing how the conclusions depend on Ng and U. I therefore recommend major revision rather than rejection, but the current form would not yet justify the strong claim that the occupation and transition energies are 'well described' by the model."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid experimental paper that shows PTCDA anion clusters can be mapped onto an extended Hubbard model with site-dependent potentials and inter-site repulsions. The new thing is the asymmetric diamond cluster: two distinct site types, distinct cross-cluster hopping terms (tAA, tBB), and site-resolved EFS maps that actually show different charge asymmetry in the equilibrium versus added/removed states. The theoretical treatment is honest and reviewable: exact diagonalization of a 4-site Hamiltonian, clear parameter constraints from geometry and prior single-molecule measurements, and the occupation asymmetry as an extra check. That is real work and the paper is readable.\n\nSoft spots: the transition energies are not genuinely predicted. The paper states that a constrained optimization algorithm was used to find V, VAA, VBB that give the correct E∆N=±1. So the agreement at 0.15 and 0.21 eV is fit-in. The occupation asymmetry then filters among the degenerate solutions; that helps, but it's qualitative because there is no quantitative conversion of surface potential to charge (the authors say so in the Discussion). Also Ng is not known experimentally; N=2 is selected because N=3 fails the constraints. So the model's central claim—that inter-site terms control the physics—rests on an inferred ground state and a qualitative charge comparison.\n\nAre these fatal? I don't think so. The model is small, exact, and the fit is transparent. The real risk is the interpretation of EFS jumps: they are assumed to be single-electron addition/removal transitions of the isolated cluster, and tip-induced charging dynamics could complicate that. A quantitative charge calibration or an independent measure of Ng would settle it. But the authors have flagged these limitations themselves, which counts for something.\n\nWho is it for? People working on molecular quantum simulators, STM/EFS charge-state readout, and small Hubbard clusters. It deserves a serious referee; the experimental data and model are specific enough that a careful referee can check the parameter logic. I'd accept it for review but would push for a clearer statement of what is fit and what is predicted, and ideally a quantitative conversion of surface potential to charge or a sensitivity analysis on Ng.","headline":"A solid experimental platform paper with a transparent model, but the key parameter extraction is fit-in and the ground-state charge number is inferred, so the central claim is conditional.","tokens_in":648,"tokens_out":1534,"would_cite":true,"duration_ms":28320,"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 four-molecule cluster of PTCDA anions on a NaCl bilayer behaves as an extended Hubbard model, with the electron addition/removal energies and the asymmetric site occupations set by inter-site repulsions and site-energy offsets rather…","keywords":["extended Hubbard model","molecular anions","PTCDA","electrostatic force spectroscopy","Hubbard clusters","electron addition/removal energies","site-resolved charge","strong correlations"],"falsifier":"Measure the EFS jump positions on the same diamond cluster while systematically varying the tip–sample distance and oscillation amplitude; genuine molecular addition/removal energies should be independent of these parameters, whereas tip-induced charging dynamics would shift the jumps with coupling strength. Alternatively, independently determine the cluster's ground-state occupation (e.g., through single-electron capacitance or a different charge-sensing method) and check whether it is indeed $N_g = 2$.","tokens_in":11740,"feed_emoji":"⚛️","tokens_out":5268,"duration_ms":46820,"temperature":0.7,"pith_summary":"The paper claims that two types of four-molecule PTCDA anion clusters—an asymmetric \"diamond\" and a symmetric \"clover\"—are quantitatively described by an extended Hubbard model. In the diamond cluster, electron addition occurs at 0.15 eV and removal at 0.21 eV, a 0.36 eV gap far below the 1.4 eV single-molecule Hubbard U, and the equilibrium charge is unevenly distributed, with the two B sites holding more negative charge than the two A sites. The model reproduces these observations only when distinct inter-site Coulomb terms ($V$, $V_{BB}$, $V_{AA}$) and site-energy offsets ($\\epsilon_A$, $\\epsilon_B$) are included, while $U$ merely prevents double occupancy. If right, this makes molecular anion clusters a tunable experimental platform for fermionic Hubbard models in a regime where non-local interactions, not $U$, control the physics.","feed_headline":"EFS jumps show inter-site forces set charge in a 4-molecule cluster","feed_subtitle":"In a PTCDA anion cluster, non-local inter-site repulsion, not Hubbard U, sets the electron addition and removal energies.","key_machinery":"The central object is the extended Hubbard Hamiltonian in Eq. (1), defined on a four-site diamond geometry. It adds to the standard nearest-neighbor hopping $t$ and on-site repulsion $U$ three types of non-local terms: diagonal hoppings $t_{BB}$ (between the two B sites) and $t_{AA}$ (between the two A sites); site-energy offsets $\\epsilon_{A/B} = \\epsilon + E_{P,A/B}$ capturing polarization differences; and inter-site Coulomb repulsions $V$ (nearest-neighbor), $V_{BB}$ (B-B diagonal), and $V_{AA}$ (A-A diagonal). The model is solved by exact diagonalization in the occupation-number basis over all spin partitions, yielding ground-state energies $E_N$ for each total occupation $N$; fitting these to the EFS jump energies and the surface-potential-derived occupations fixes $V$, $V_{BB}$, $V_{AA}$, and $N_g = 2$. The load-bearing mechanism is that these non-local terms lower the energy of specific charge configurations, thereby controlling both the transition energies and the site-resolved partial occupations, with $U$ only forbidding double occupancy.","core_discovery":"The central claim is that in the large-$U$ limit ($t \\ll U$), the charge-state transitions and occupation asymmetry of the diamond cluster are governed by the extended terms of the Hubbard Hamiltonian, not by $U$ itself. With $U = 1.4$ eV, $\\epsilon_B = 0.485$ eV, $\\epsilon_A = 0.462$ eV, $V = 0.315$ eV, $V_{BB} = 0.281$ eV, $V_{AA} = 0.247$ eV, and $t = 20$ meV (with $t_{AA} = t_{BB} = 0$), the model matches the measured single-electron addition (0.15 eV) and removal (0.21 eV) energies and the site-resolved surface-potential maps for $N = 1$, $2$, $3$ charge states, assuming a ground-state occupation $N_g = 2$. Notably, the occupation asymmetry reverses between $N = 2$ (B sites more occupied) and $N = 3$ (A sites more occupied), a feature driven by the cross-cluster hoppings and inter-site potentials.","pith_inferences":["If the $N_g = 2$ assignment is confirmed, the diamond cluster is effectively a two-electron (or two-hole) plaquette; the same geometry might exhibit spin correlations or charge ordering that could be probed with spin-sensitive scanning probes, a direction the paper mentions but does not pursue.","The predicted reversal of occupation asymmetry between $N = 2$ and $N = 3$ (B sites more occupied vs. A sites more occupied) is a sharp, testable fingerprint; a quantitative conversion of measured surface potential to absolute charge could confirm it directly.","For larger arrays of such anion clusters, the EFS technique could map the boundary between charge-ordered and delocalized phases as a function of $V/U$ and site-energy disorder, effectively realizing a digital-twin experiment for extended Hubbard phase diagrams.","The result implies that in this platform the intersite repulsions act as the dominant energy scale for charge dynamics, so cluster-based simulators may need to include these terms explicitly rather than treating the Hubbard model as the effective theory."],"forward_implications":["The same exact-diagonalization-plus-EFS approach can be applied to larger and differently shaped molecular clusters, allowing experimental exploration of a wider phase space of extended Hubbard models.","The inter-site potentials and site-energy offsets are tunable through molecule choice, substrate (e.g., different ionic salts), and film thickness, so the ratios $V/U$ and $\\epsilon/U$ could be varied systematically.","The demonstration that non-local terms, not $U$, control charge-state transitions in the strongly localized limit suggests that small molecular clusters can reveal correlation physics missed by the single-band Hubbard model alone.","Site-resolved EFS maps provide a direct readout of partial occupations for each charge state, enabling quantitative comparison with exact-diagonalization predictions across a range of parameters."],"supporting_citations":[{"why":"Demonstrates that electrostatic force spectroscopy features correspond to single-electron addition/removal energies of a few-electron quantum dot, providing the interpretive basis for the EFS jumps used here.","marker":"[17]"},{"why":"Supply the PTCDA-on-NaCl electronic structure, the singly occupied molecular orbital, and the polarization-induced energy-level shifts that set the site-energy offsets $\\epsilon_A$ and $\\epsilon_B$.","marker":"[28]"},{"why":"Shows how pixel-by-pixel electrostatic force imaging maps the local charge distribution within a single molecule, which justifies using surface-potential maps as proxies for site occupations.","marker":"[21]"},{"why":"Establishes that on insulating multilayer films, the charge state of individual adsorbates can be switched and stabilized, supporting the assignment of the EFS jumps as transitions of the cluster's charge state.","marker":"[23]"},{"why":"Models charge-state dynamics in electrostatic force spectroscopy, addressing the relation between transient charging events and the measured frequency-shift jumps.","marker":"[24]"}],"fun_headline_variants":["Inter-site forces, not U, set charge in molecular cluster","Extended Hubbard terms dictate electron addition in 2D anion cluster","Charge transitions in PTCDA cluster driven by non-local repulsion","Molecular anion cluster reveals extended Hubbard model physics","In PTCDA cluster, inter-site repulsion outshines Hubbard U"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the EFS frequency-shift jumps directly equal the single-electron addition and removal energies of an isolated cluster, and that the cluster's ground state holds exactly two electrons ($N_g = 2$); the paper states that \"the value of $N_g$ is not known from the experiment\" and selects $N_g = 2$ because $N_g = 3$ fails the imposed occupation and gap constraints.","fun_headline_variants_meta":{"raw":{"variants":["Inter-site forces, not U, set charge in molecular cluster","Extended Hubbard terms dictate electron addition in 2D anion cluster","Charge transitions in PTCDA cluster driven by non-local repulsion","Molecular anion cluster reveals extended Hubbard model physics","In PTCDA cluster, inter-site repulsion outshines Hubbard U"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001175,"raw_usage":{"total_tokens":4878,"prompt_tokens":987,"completion_tokens":3891,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":603,"completion_tokens_details":{"reasoning_tokens":3805}},"tokens_in":603,"tokens_out":3891,"duration_ms":26577,"temperature":1.0,"reasoning_tokens":3805,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:20:26.269889+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the EFS jump positions on the same diamond cluster while systematically varying the tip–sample distance and oscillation amplitude; genuine molecular addition/removal energies should be independent of these parameters, whereas tip-induced charging dynamics would shift the jumps with coupling strength. Alternatively, independently determine the cluster's ground-state occupation (e.g., through single-electron capacitance or a different charge-sensing method) and check whether it is indeed $N_g = 2$.","supporting_citations":[{"cited_title":"Cockins, Y","cited_arxiv_id":null,"evidence_quote":"Demonstrates that electrostatic force spectroscopy features correspond to single-electron addition/removal energies of a few-electron quantum dot, providing the interpretive basis for the EFS jumps used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supply the PTCDA-on-NaCl electronic structure, the singly occupied molecular orbital, and the polarization-induced energy-level shifts that set the site-energy offsets $\\epsilon_A$ and $\\epsilon_B$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how pixel-by-pixel electrostatic force imaging maps the local charge distribution within a single molecule, which justifies using surface-potential maps as proxies for site occupations."},{"cited_title":"Steurer, J","cited_arxiv_id":null,"evidence_quote":"Establishes that on insulating multilayer films, the charge state of individual adsorbates can be switched and stabilized, supporting the assignment of the EFS jumps as transitions of the cluster's charge state."},{"cited_title":"Ondr´ aˇ cek, P","cited_arxiv_id":null,"evidence_quote":"Models charge-state dynamics in electrostatic force spectroscopy, addressing the relation between transient charging events and the measured frequency-shift jumps."}],"review_version":2}