{"id":"536665eb-1f19-44d4-b3a5-3d84764478c6","arxiv_id":"1908.06580","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Two coupled graphene quantum dots form bonding and antibonding quasibound molecular states, directly imaged by STS, and magnetic fields lift their angular momentum degeneracy.","lead":"Two coupled circular graphene quantum dots form bonding and antibonding quasibound states, which the authors image directly with a scanning tunneling microscope. The result extends the quantum-dot artificial molecule concept from ordinary electrons to massless Dirac fermions in graphene.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The molecular-state assignment depends on two nearly degenerate isolated QDs; a 30 meV doublet could equally arise from a size mismatch or disorder, and the reported LDOS maps are not subjected to a quantitative symmetry test.","rationale":"The reader's weakest_assumption identifies the same vulnerability I see, and I agree with it. I considered three possible objections: the g* ≈ 40 fit lacking error bars; the magnetic-field data coming from a single dot; and the possibility that the 30 meV doublet is a detuning or disorder effect rather than hybridization. The g* issue is real but secondary because the molecular-state claim does not depend on the exact g value. The single-dot magnetic-field data are a reproducibility concern, not a logical flaw in the central interpretation. The detuning/disorder objection is load-bearing because every observable used to support the molecular state, namely the doublet spacing and the two LDOS maps, is also what one would expect, with suitable parameters, from two slightly different dots or from disorder-broken degeneracy in a single dot. The paper's own energy relation makes this alternative quantitatively plausible: a 25% radius mismatch would produce a 30 meV level shift, and such a mismatch is not quantified or excluded. The lattice Green's function calculation is genuine independent support, but it uses idealized identical circles and cannot by itself establish that the experimental pair realizes that limit. I therefore do not reject the paper; the interpretation is plausible, and the magnetic-field evolution is a useful additional check. However, the central claim remains conditional on the dot-equivalence and symmetry checks that the paper does not report. Since the reader's CONDITIONAL verdict already reflects that conditionality, I recommend UNCHANGED.","tokens_in":8100,"tokens_out":5521,"duration_ms":64714,"concrete_test":"Digitize the raw STS maps at the two peak energies (Figs. 2(f) and 2(g)) and integrate the LDOS in the left dot, the right dot, and the inter-dot barrier. Compute an asymmetry score A = (LDOS_right − LDOS_left)/(LDOS_right + LDOS_left) for each peak and check for a nodal minimum along the perpendicular bisector for the upper peak and enhanced barrier LDOS for the lower peak. For bonding/antibonding states of nearly identical dots, both peaks should have |A| near 0; if instead each peak is localized mainly in one dot (|A| near 1), the doublet is consistent with two uncoupled dots having a ~30 meV level detuning rather than with hybridization.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the observed 30 meV doublet is a bonding/antibonding splitting of the lowest quasibound state of two coupled graphene QDs. That inference requires the uncoupled levels of the two dots to be nearly degenerate and the double-peak spatial maps to be the symmetric and antisymmetric combinations. The paper supports degeneracy only by saying the two QDs 'almost have identical size and structure' and by showing that their spectra are similar; no measured radii, area difference, or error bars are given. Using the authors' own estimate ΔE ≈ ħv_F/R, a 30 meV detuning between uncoupled dots would require only δR/R ≈ 30/120 ≈ 25%, a difference that could easily escape visual inspection in one STM image. Conversely, if the two dots are indeed identical, the 30 meV splitting could still be an intrinsic single-dot feature, for example a disorder- or tip-induced splitting of the non-WGM lowest state; the only comparison is to a different isolated dot in the supplementary material, not to a control dot of the same size and environment with coupling turned off. The LDOS maps are asserted to 'directly demonstrate' bonding and antibonding states, but no quantitative nodal-plane or symmetry analysis is presented, and the caption reports maps recorded in the left dot only. Because the strongest evidence (Figs. 2(d)–2(g)) comes from a single pair of dots without statistics or a symmetry test, the molecular-orbital interpretation is not uniquely determined.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports STM/STS experiments on two coupled circular graphene quantum dots (QDs) formed by S nanoclusters underneath a graphene monolayer on Cu. The authors observe a ~30 meV splitting of the lowest quasibound state in the two coupled QDs, image two distinct LDOS maps that they assign to bonding (σ) and antibonding (σ*) molecular states, and find in magnetic fields that each molecular state further splits into two peaks with a linear field dependence, yielding an effective orbital g factor g* ≈ 40. These observations are compared with lattice Green's function calculations for two coupled circular QDs and with WKB calculations for the magnetic-field splitting. The paper concludes that this constitutes the first realization of a relativistic artificial molecule made of massless Dirac fermions.","tokens_in":8458,"tokens_out":3999,"duration_ms":40843,"significance":"If the central claims hold, this is a notable advance: it would extend the artificial-molecule paradigm, previously limited to nonrelativistic fermions, to massless Dirac fermions in graphene QDs, and it would directly visualize bonding and antibonding states of such a molecule. The paper also reports a magnetic-field-induced lifting of orbital degeneracy in the molecular states, which is an interesting and falsifiable result. Strengths of the manuscript include the use of an independent lattice Green's function calculation that yields a split lowest quasibound state, the energy-dependent reduction of the splitting with increasing state energy (consistent with WGM confinement), and the direct real-space STS maps. However, the quantitative support is thin: no error bars are given for the splitting or the g* fit, the molecular-state assignment relies on a single pair of QDs without statistical replication, and the spatial maps are not subjected to a quantitative symmetry test. These issues are load-bearing for the main claim, so the paper requires substantial revision.","major_comments":[{"comment":"The assignment of the ~30 meV doublet to bonding/antibonding molecular states requires that the two uncoupled QDs have nearly degenerate lowest quasibound levels. The only support given is the statement that the two QDs 'almost have identical size and structure' and that their spectra are 'almost the same'; no measured radii, area difference, or error bars on peak positions are provided. Using the authors' own estimate ΔE ≈ ħv_F/R ≈ 120 meV, a level detuning of 30 meV between uncoupled dots would require only a radius mismatch δR/R ≈ 25%, a difference that could easily escape visual inspection in a single STM image. Without a control experiment (e.g., a single QD of the same size and environment, or a pair with coupling intentionally varied) or a quantitative comparison of the two dots' dimensions and spectra, the doublet could equally arise from dot-to-dot variation or disorder. This is the central premise of the molecular-state claim and needs direct quantitative support.","section":"Figure 2(c)-2(e) and surrounding text"},{"comment":"The LDOS maps are stated to 'directly demonstrate' the formation of bonding and antibonding states, but they were recorded only in the left QD, and no quantitative symmetry or nodal-plane analysis is presented. A bonding state should be symmetric about the inter-dot axis and an antibonding state antisymmetric (or vice versa depending on the basis), but the manuscript does not show that the measured maps possess the required symmetry, nor does it compare them quantitatively with the calculated LDOS in Figs. 3(c) and 3(d). Without such a test, the maps do not uniquely determine the molecular-orbital assignment; a single-dot asymmetric state or a tip-induced artifact could produce similar-looking spatial patterns.","section":"Figures 2(f)-2(g) and 3(c)-3(d)"},{"comment":"The effective g factor g* ≈ 40 is extracted from a linear fit to the magnetic-field splitting, but no error bars, number of data points, or reproducibility across different dot pairs are reported. Furthermore, the WKB calculation explicitly assumes constant inter-dot coupling, and the model parameters (R = 6 nm, d = 4 nm, ΔV = 270 meV) are chosen ad hoc; no sensitivity analysis for these parameters is provided. The statement that the experimental values are 'slightly larger' than the theoretical calculation is only qualitative. These omissions limit the quantitative reliability of the large orbital g factor claim.","section":"Figure 4 and text on g*"},{"comment":"The claim that 'the fourfold degeneracy of the first quasibound state also proves the formation of the artificial molecule' is not justified. The number of observed peaks in a magnetic field depends on the single-dot degeneracy structure (including orbital m, spin, and valley degeneracies) as well as on the inter-dot coupling; four peaks alone do not constitute a proof of molecular formation. A detailed level-counting and symmetry argument is needed before this statement can support the molecular-state assignment.","section":"Magnetic-field degeneracy argument"}],"minor_comments":[{"comment":"The sentence 'on top of the ordered S superlattice formed advanced' appears garbled; 'formed advanced' should likely be 'formed beforehand' or similar.","section":"Full text near 'S superlattice'"},{"comment":"'consisting with that of monolayer S atoms' should be 'consistent with that of monolayer S atoms'.","section":"Full text, 'consisting with'"},{"comment":"The caption uses 'red cycle' and 'dash squares'; these should read 'red circle' and 'dashed squares'.","section":"Figure 4 caption"},{"comment":"The symbols σ*_+, σ*_-, σ_+, σ_- are used in the text and figure but are not explicitly defined in the main text; the definition as the ±m sublevels of the bonding and antibonding states should be stated in the caption or main text.","section":"Figure 4(a) markers"},{"comment":"The phrase 'the split of the quasibound states decreases' would be clearer as 'the splitting of the quasibound states decreases'.","section":"Text near Fig. 2(c)"},{"comment":"The main text repeatedly refers to supporting materials for critical data (e.g., isolated-dot control, FER measurements, Landau-level Dirac point); the key quantitative results from those sections should be summarized in the main text so the claims can be evaluated without accessing the SI.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a conceptually appealing experiment, but the evidence for the central molecular-state interpretation is not yet quantitatively compelling. The reliance on a single dot pair, the absence of error bars and control measurements, and the unquantified symmetry of the LDOS maps are the main concerns. I would encourage the authors to provide the missing statistical and control data, or to temper the strength of the claims accordingly. The theoretical Green's function calculation is a positive element and should be retained."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe one thing to know: this paper is the first direct STM/STS study of two coupled graphene quantum dots, and it claims to see bonding/antibonding molecular states. On the merits, the claim is plausible and the data are consistent, but it rests on a single pair of dots and on an assumption of near-identical dots that is not quantified. The risk is real: a 30 meV doublet could also arise from a ~25% size mismatch or disorder between the two dots, since the level spacing itself is ~120 meV.\n\nWhat's genuinely new: previous artificial molecules used nonrelativistic carriers; this is the first coupled graphene QD system, where the carriers are massless Dirac fermions. The observed splitting of the lowest quasibound state into two peaks, with the splitting decreasing at higher energy, matches the expectation that the non-WGM lowest state couples most. The lattice Green's function calculation reproduces the main features without fitting the splitting itself, so the molecular interpretation is not obviously circular. The magnetic-field data showing each molecular state splitting into ±m sublevels is a nice extra, and the extracted g* ≈ 40 is a meaningful orbital moment, even without error bars.\n\nSoft spots: the central inference depends on the two dots being effectively degenerate when uncoupled. The paper says they \"almost have identical size and structure\" but gives no measured radii, area difference, or error bars. The control comparison is to an isolated dot elsewhere, not to a dot of the same size with coupling disabled. The LDOS maps are recorded in the left dot only, and the claim of bonding/antibonding symmetry is backed by eye rather than by a nodal-plane analysis. All of this is soft, not fatal: the spectra look reasonable, and the theory supports the assignment. But it means the paper is a compelling initial observation, not a bulletproof one.\n\nWho should read it: anyone working on graphene quantum confinement, Klein tunneling, or artificial molecules. It deserves a serious referee; I'd send it to review with a request for more statistics and a quantitative symmetry check, but I wouldn't desk reject it.\n\nLet me know if you want to discuss.","headline":"Plausible first observation of a graphene quantum dot molecule, but the central splitting needs more control data before I'd call it definitive.","tokens_in":8946,"tokens_out":2264,"would_cite":false,"duration_ms":24337,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Coupled graphene dots form a relativistic artificial molecule","keywords":["graphene quantum dots","artificial molecules","relativistic Dirac fermions","bonding and antibonding states","scanning tunneling spectroscopy","whispering-gallery modes","Klein tunneling","orbital g-factor"],"falsifier":"Measure tunneling spectra on a single isolated graphene dot made by the same sulfur-island method; if its lowest quasibound state also shows two peaks, the splitting is an intra-dot or tip effect rather than inter-dot bonding, and in a coupled pair the splitting should shrink as the inter-dot separation increases.","tokens_in":7937,"feed_emoji":"⚛️","tokens_out":12294,"duration_ms":112005,"temperature":0.7,"pith_summary":"The paper reports that two coupled circular graphene quantum dots behave as a relativistic artificial molecule, a molecule analogue whose electrons are massless Dirac fermions rather than ordinary Schrödinger electrons. It shows, with scanning tunneling spectroscopy, that the lowest quasibound state of the pair splits into a bonding and an antibonding state separated by about 30 meV, and that spatial maps of the local density of states directly image both molecular states. The same relativistic character makes the molecular levels respond to a magnetic field by lifting the $\\pm m$ angular-momentum degeneracy, so each molecular peak splits into two with an effective orbital $g$-factor of about 40. The result is presented as the first coupled graphene quantum-dot molecule and a platform in which molecular orbitals are built from confined massless Dirac fermions.","feed_headline":"Coupled graphene dots form a relativistic artificial molecule","feed_subtitle":"STM maps reveal bonding and antibonding states; magnetic fields split them by angular momentum.","key_machinery":"The carrying object is the circular graphene p-n junction as a whispering-gallery cavity for massless Dirac fermions. A nanoscale sulfur island raises the local Dirac point by about 210 meV relative to the surrounding graphene, and Klein tunneling across the resulting p-n boundary traps electrons in quasibound states. In two coupled dots the molecular states are formed by the sum and difference of the individual dot wavefunctions, $\\psi_\\sigma \\approx \\psi_L + \\psi_R$ and $\\psi_{\\sigma^*} \\approx \\psi_L - \\psi_R$. The coupling is strongest for the lowest quasibound state because its long wavelength makes it least confined, so the splitting decreases with increasing energy. The theoretical support comes from lattice Green's function calculations of two circular QDs of radius 6 nm, separation 4 nm, and potential step 270 meV, and for the magnetic response a WKB treatment with fixed inter-dot coupling.","core_discovery":"On the paper's own terms, the central discovery is the hybridization of confined Dirac-fermion states across two adjacent graphene quantum dots. Each dot is a circular p-n junction formed by a nanoscale sulfur island between the graphene and the copper substrate; massless Dirac fermions are temporarily trapped in whispering-gallery quasibound states. When two such dots sit close together, the lowest quasibound state, the one least protected by whispering-gallery confinement, couples most strongly and splits into a lower bonding state and a higher antibonding state with a measured separation of about 30 meV. STS maps at the two peak energies show the two molecular orbitals, and theoretical lattice Green's function calculations reproduce both the level splitting and the spatial distributions. In a magnetic field each molecular level further splits linearly with field into opposite-angular-momentum sublevels, and fitting $\\Delta E = g^*\\mu_B B$ gives $g^* \\approx 40$, indicating an orbital rather than spin origin. The authors take these observations as direct evidence that relativistic artificial molecules exist and behave qualitatively differently from their nonrelativistic counterparts.","pith_inferences":["Beyond the paper: if the molecular-orbital assignment is correct, varying the inter-dot separation while holding dot size fixed should tune the 30 meV splitting, providing a control experiment the current geometry does not perform.","Beyond the paper: the same fabrication and imaging route could be extended to three or more coupled dots, where the hybridization should form one-dimensional relativistic molecular bands.","Beyond the paper: because the measured $g^* \\approx 40$ is orbital in origin, Dirac-dot-molecule devices aimed at spin-based quantum information would need to account for field-induced orbital level shifts rather than spin splitting alone.","Beyond the paper: the linear field dependence of the $\\pm m$ splitting could serve as a local probe of the p-n junction potential profile, since sharper confinement should modify the orbital moment."],"forward_implications":["The 30 meV bonding-antibonding splitting should appear as a robust double-peak feature in STS taken at the center of either dot.","The spatial contrast between bonding and antibonding STS maps provides a real-space fingerprint for identifying molecular states in coupled graphene dots.","Magnetic fields should split each molecular peak into two sublevels whose separation grows linearly with field while the bonding-antibonding separation stays nearly constant.","The effective orbital $g$-factor near 40 means modest magnetic fields can substantially retune the level structure of graphene-dot molecules.","Because coupling weakens for higher angular-momentum quasibound states, only the low-lying levels of coupled graphene dots are expected to show clear molecular hybridization."],"supporting_citations":[{"why":"Establishes whispering-gallery-mode confinement of massless Dirac fermions in graphene quantum dots, the confinement mechanism extended here to coupled dots.","marker":"[26]"},{"why":"Shows Klein tunneling and electron trapping in nanoscale graphene quantum dots, the physical basis for the quasibound states.","marker":"[27]"},{"why":"Demonstrates imaging of confined Dirac fermions in graphene quantum dots, the STS mapping approach used to visualize the molecular states.","marker":"[28]"},{"why":"Provides the lattice Green's function method for graphene p-n junctions used in the coupled-dot calculations.","marker":"[29]"},{"why":"Gives the chiral-tunneling (Klein paradox) theory that explains transmission across the graphene p-n boundaries.","marker":"[30]"},{"why":"Supplies the sulfur-nanocluster fabrication route that creates the circular p-n junctions forming the quantum dots.","marker":"[31]"},{"why":"Numerically studies Klein quantum dots in graphene and underlies the calculated LDOS of the coupled system.","marker":"[35]"},{"why":"Explains magnetic-field-induced lifting of the $\\pm m$ degeneracy in Dirac quantum dots via Berry-phase jumps, the framework for the observed field splitting.","marker":"[41]"}],"fun_headline_variants":["Graphene dots hybridize into relativistic molecule","Relativistic molecule from coupled graphene quantum dots","Bonding and antibonding in relativistic graphene molecule","Magnetic field splits relativistic graphene molecular states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the two dots are nearly identical and that the observed 30 meV splitting comes from inter-dot coupling rather than from differences in dot size, disorder, or tip-induced effects.","fun_headline_variants_meta":{"raw":{"variants":["Graphene dots hybridize into relativistic molecule","Relativistic molecule from coupled graphene quantum dots","Bonding and antibonding in relativistic graphene molecule","Magnetic field splits relativistic graphene molecular states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000237,"raw_usage":{"total_tokens":1498,"prompt_tokens":930,"completion_tokens":568,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":508}},"tokens_in":546,"tokens_out":568,"duration_ms":6658,"temperature":1.0,"reasoning_tokens":508,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:40:38.339205+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure tunneling spectra on a single isolated graphene dot made by the same sulfur-island method; if its lowest quasibound state also shows two peaks, the splitting is an intra-dot or tip effect rather than inter-dot bonding, and in a coupled pair the splitting should shrink as the inter-dot separation increases.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes whispering-gallery-mode confinement of massless Dirac fermions in graphene quantum dots, the confinement mechanism extended here to coupled dots."},{"cited_title":"Gutié rrez, L","cited_arxiv_id":null,"evidence_quote":"Shows Klein tunneling and electron trapping in nanoscale graphene quantum dots, the physical basis for the quasibound states."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates imaging of confined Dirac fermions in graphene quantum dots, the STS mapping approach used to visualize the molecular states."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the lattice Green's function method for graphene p-n junctions used in the coupled-dot calculations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the chiral-tunneling (Klein paradox) theory that explains transmission across the graphene p-n boundaries."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the sulfur-nanocluster fabrication route that creates the circular p-n junctions forming the quantum dots."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Numerically studies Klein quantum dots in graphene and underlies the calculated LDOS of the coupled system."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Explains magnetic-field-induced lifting of the $\\pm m$ degeneracy in Dirac quantum dots via Berry-phase jumps, the framework for the observed field splitting."}],"review_version":1}