{"id":"83501968-b93d-4e36-ae73-72c2d3552468","arxiv_id":"2411.18785","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A four-layer graphene double-moiré device allows independent tuning and thermodynamic probing of two twisted-bilayer-graphene flat bands, revealing correlated insulators near magic angle and charge-neutrality gaps at larger twist angles.","lead":"Researchers built a four-layer graphene stack containing two twisted bilayer graphene devices, one on top of the other. They can tune each flat band independently and measure how electron interactions in each band change with twist angle.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The two-TBG thermodynamic analysis assumes separate layer chemical potentials and conserved carrier numbers, but the edge contacts to the four-layer stack likely short the layers, invalidating Eq. (1) and the extracted μ(n) curves.","rationale":"The reader's verdict was CONDITIONAL and identified the separate-conservation assumption as the weakest. I agree that this is the most load-bearing assumption, and I sharpen it: the failure mode is not only interlayer tunneling but the electrical short imposed by edge contacts to the full stack. This matters because the paper's central novelty is the thermodynamic extraction of independent μ(n) for each TBG; if the two subsystems cannot sustain different electrochemical potentials, that extraction is invalid even though the qualitative existence of two moiré flat bands may survive. The neutrality-gap ambiguity and fitted theoretical parameters are secondary; they can be fixed by rewording or more data, whereas the two-chemical-potential framework is foundational. The proposed test would settle the issue by comparing the data with a common-μ model. Therefore I do not change the reader's CONDITIONAL verdict; the paper needs a direct demonstration that the two TBGs are electrically isolated (e.g., independent contacts or interlayer resistance measurement), or an acknowledgment that the layer-resolved thermodynamic quantities assume an isolation that the device geometry does not guarantee.","tokens_in":16308,"tokens_out":14438,"duration_ms":139247,"concrete_test":"Re-derive the gate-voltage relations for the four-layer stack under the assumption of a single common electrochemical potential (no independent μ_T, μ_B), and fit the incompressible-state diamond slopes in Fig. 2(d) and the Landau-fan capacitance ratios in Fig. S2 with both models. If the common-μ model reproduces the data without invoking C_IL, the separate-conservation foundation of Eq. (1) is not supported and the extracted layer-resolved chemical potentials are model artifacts.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation (1) of the main text models the double moiré system as two independent 2D electron gases with separately conserved particle numbers and distinct chemical potentials μ_T, μ_B, coupled only by an interlayer capacitance C_IL. This ansatz requires the top and bottom TBG subsystems to be electrically isolated from each other except for capacitive coupling. The device, however, is a single four-layer stack etched into a Hall bar with edge contacts deposited on the side of the stack; such contacts necessarily touch all four graphene layers, shorting the two TBG subsystems together. In a contacted stack, the electrochemical potential is common to all layers in the absence of interlayer current, so μ_T = μ_B and the independent-conservation model in Eq. (1) is not the correct starting point. The standard description is a dual-gated conductor with total density and displacement field as independent variables; pinning of the common Fermi level when one TBG is incompressible can create resistance features that resemble those interpreted as independent layer fillings. The consistency check reported after Eq. (3)—identical μ_B vs. n_B at different integer ν_T—does not rule out a common-μ model, since a pinned Fermi level would give the same apparent bottom-layer relation while the top-layer density changes. The extracted C_IL = 1.6 μF/cm² and the layer-resolved μ(n) curves in Fig. 3 are therefore not independently validated, and the central thermodynamic claim is at risk.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports transport measurements on a double moiré system of four graphene layers, in which the top and bottom pairs form twisted bilayer graphene (TBG) with small twist angles and the middle interface has a large rotational mismatch. The authors observe resistance peaks in the (VTG, VBG) plane that they attribute to integer fillings of the top and bottom TBG flat bands. Using a thermodynamic model of two capacitively coupled layers with separately conserved densities (Eq. 1), they extract chemical potential versus density curves for each TBG, identify correlated insulators near half and quarter fillings, and report a twist-angle dependence of gaps at neutrality and at integer fillings. They compare the extracted gaps with self-consistent Hartree calculations and find qualitative agreement after adjusting phenomenological parameters. The paper concludes that the two flat bands are independently tunable, and that valence flat bands are flatter than conduction flat bands.","tokens_in":16545,"tokens_out":13840,"duration_ms":128972,"significance":"The paper proposes a new architecture for studying two moiré flat bands in one device and provides a data set spanning twist angles from 0.91° to 1.57°. The temperature dependence in Fig. 4 confirms that some observed resistance peaks are correlated insulators, and the comparison with previous chemical potential measurements of TBG is valuable. However, the central quantitative claim of layer-resolved chemical potentials rests on the assumption of separately conserved densities in the two TBG subsystems, which is undermined by the edge-contact geometry. The Hartree comparison is a parameter fit rather than a prediction. If the contact issue could be resolved or the analysis redone under a common-μ model, the empirical observations might still be of interest, but as presented the main claim is not established.","major_comments":[{"comment":"The thermodynamic model in Eq. (1) assumes that the top and bottom TBG subsystems have separately conserved carrier numbers and are coupled only by an interlayer capacitance C_IL. This assumption is not justified by the device geometry. The samples are four-layer stacks etched into a Hall bar with edge contacts, which contact all graphene layers and thus short the two TBG subsystems at the contacts. Since the channel is shorted to ground (SM Sec. I), the electrochemical potentials of the two subsystems are equal at the contacts and charge can be exchanged between them through the contacts. Consequently, n_T and n_B are not independently conserved, and Eq. (1) is not the correct starting point; a dual-gated conductor with total density and displacement field as independent variables is the standard description. The consistency check after Eq. (3), namely identical μ_B vs n_B at different integer ν_T, does not rule out a common-μ model, because a pinned Fermi level would yield the same apparent bottom-layer relation while the top-layer density changes. This issue directly affects the central claim of measuring layer-resolved chemical potentials and the assignment of resistance peaks to independent layer fillings.","section":"Eq. (1) and SM Sec. I"},{"comment":"The comparison with the self-consistent Hartree approximation is presented as a validation, but the model parameters are set to match the data. As stated in the text, α is set to 0.3 for θ=0.91°, 0.6 for θ=0.99°–1.57°, and 1.0 for θ=1.7° specifically to reproduce the measured Eg,±1, and wNL is arbitrarily set to −20 meV. The paper itself notes that this choice is not a rigorous determination. Therefore, the agreement in Fig. 3(c)-(d) is a fit, not a parameter-free prediction, and it does not provide independent confirmation of the extracted gap sizes or their twist-angle dependence. The manuscript should either constrain α and wNL from independent data or present the curves as fits with an explicit discussion of the parameter uncertainty.","section":"Fig. 3(c)-(d) and SM Sec. V"},{"comment":"The extracted chemical potential curves, gap sizes Eg,ν, and bandwidth measures Δμp and Δμn are reported without error bars or uncertainty propagation. The capacitance values CTG, CBG, and CIL enter directly into Eqs. (2) and (3), and the extraction relies on assumptions such as Δμ_X << eΔV in SM Eq. (S5). The sample-to-sample scatter in Fig. 3 is significant, so without a quantitative uncertainty estimate the claimed twist-angle trends, including the minimum of Eg,0 near the magic angle, are not firmly established. The authors should provide error bars derived from capacitance uncertainties and sample-to-sample variations.","section":"Figs. 3(b)-(d) and SM Sec. II"}],"minor_comments":[{"comment":"SM Sec. III refers to 'Fig. S7(d)' for the boundary of the νT = 0 state, but Fig. S7 is the µ vs n plot; the intended figure appears to be Fig. S5(d).","section":"SM Sec. III"},{"comment":"The main text contains typos, e.g., 'exprimental' in the caption of Fig. 3, and the manuscript would benefit from a careful proofread.","section":"General"},{"comment":"In Fig. 3(d), the scattered literature data are not fully identified in the caption; specify which symbols correspond to Refs. 27, 29, 30, and 32.","section":"Fig. 3(d)"},{"comment":"The discussion of superconductivity uses the observed electron-hole asymmetry to speculate that flatter valence bands favor superconductivity, but no superconductivity is observed in this study; this statement should be labeled as a conjecture rather than a conclusion of the present data.","section":"Last paragraph"}],"recommendation":"reject","confidential_remarks":"The device geometry—edge contacts to a four-layer stack—makes it very likely that the two TBG subsystems are shorted at the contacts, invalidating the separate-density model. The consistency check in the paper does not discriminate between the two models. This is a fundamental issue that cannot be fixed by a revision of the present data; new devices with independently contacted TBG layers or a full re-analysis under a common-μ model would be required. I therefore recommend rejection, despite the interesting data set."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Luis, quick take on arXiv:2411.18785. The device is new and the experiment is careful: a four-layer stack with two small-twist TBG subsystems separated by a large-mismatch interface, and they extract μ(n) for each subsystem over a 0.91–1.57° range. The transport maps in Fig. 1 and the temperature dependence of the correlated insulators in Fig. 4 are convincing. The observation that quarter-filling gaps are strongest near magic angle while the neutrality dips get larger away from it is a real dataset that anyone building flat-band models will want.\n\nThe method is an adaptation of the double-layer thermodynamic technique from Refs. 25–28, applied to two TBGs. The consistency check after Eq. (3) — identical μ_B(n_B) at different ν_T — is a reasonable support for weak interlayer coupling, though it is not a proof. The Hartree comparison in Fig. 3(c)-(d) is explicitly a fit: α is chosen per twist angle and w_NL is set to a constant. The authors say so plainly, so this is not a hidden fit, but it means the theory does not independently predict the trends. The charge-neutrality “gap” is handled honestly: they note it could be a diverging Fermi velocity and state that compressibility measurements cannot easily distinguish the two. That is the right thing to say, but it weakens the abstract’s claim about gapped states at neutrality.\n\nOn the stress-test: the concern that edge contacts short the layers and invalidate Eq. (1) does not actually hold up on reading. The electrostatic derivation in the SM includes the equilibrium condition (0 = -eV_ε,IL + μ_B - μ_T) which enforces equal electrochemical potentials in the two layers. That is exactly what a shorted contact imposes. The phrase “separately conserved particle numbers” invites trouble, but the equations are the standard dual-gate bilayer electrostatics with no interlayer tunneling. The real assumption is the absence of tunneling/hybridization across the middle interface, which the large angle and the consistency check support. The authors should rewrite that sentence and explicitly address why edge contacts do not change the analysis, because as written it will draw referee fire.\n\nWeaknesses are mostly missing error bars and the lack of an independent check of the layer-resolved densities. The C_IL extraction from diamond geometry is clever but would benefit from a fuller uncertainty discussion. The null result on superconductivity is fine, but the speculative paragraph on flatness and superconductivity is a bit of a stretch.\n\nBottom line: this deserves serious peer review. It is a high-quality experimental paper with a new platform and a useful dataset. I would send it to referees, asking for a clearer statement of the electrochemical model, error bars on the extracted gaps, and a toned-down neutrality-gap claim in the abstract.","headline":"A new double-moiré platform with credible independent flat bands; the thermodynamic extraction is plausible, but the separate-layer model needs a clearer defense and the theory comparison is explicitly a fit.","tokens_in":17173,"tokens_out":7781,"would_cite":true,"duration_ms":72680,"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-layer graphene stack can host two spatially separated twisted-bilayer flat bands, and a capacitance model lets each band's chemical potential be measured independently.","keywords":["double moiré system","twisted bilayer graphene","flat bands","chemical potential","correlated insulator","charge neutrality gap","magic angle","thermodynamic measurement"],"falsifier":"A decisive test would be a double-moiré device with the middle twist angle deliberately reduced below about 5°: if interlayer tunneling matters, constant-$\\mu$ traces from the top layer would no longer give identical $\\mu_{\\rm B}$ versus $n_{\\rm B}$ curves at different top-layer fillings, and the diamond-edge slopes would deviate from the capacitance-only prediction. Alternatively, Landau-fan spectroscopy at matched densities could reveal tunnel-induced anticrossings between the two subsystems.","tokens_in":16054,"feed_emoji":"🔬","tokens_out":11665,"duration_ms":168391,"temperature":0.7,"pith_summary":"This paper reports that a stack of four graphene layers, with a small twist angle between the top pair and between the bottom pair but a deliberately large twist angle between the two pairs, forms two separated twisted-bilayer-graphene (TBG) flat-band systems in a single device. Using a dual-gated channel and an electrostatic model in which the two TBG subsystems share only an interlayer capacitance, the authors extract the chemical potential as a function of carrier density for each TBG across twist angles from 0.91° to 1.57°. They find that correlated insulating states at integer numbers of electrons per moiré unit cell are strongest for twists near the magic angle, while gapped states at charge neutrality are more robust at larger twist angles, and that the valence flat band is flatter than the conduction band. The result opens a way to study flat-band correlations while keeping two flat bands independently tunable in one stack.","feed_headline":"A four-layer stack hosts two independently tunable flat bands","feed_subtitle":"Each twisted bilayer in the stack is tunable and measurable alone, exposing how correlation gaps depend on twist angle.","key_machinery":"The central mechanism is the two-capacitor electrostatic model of the double moiré system, written as Eq. (1): $V_{\\rm BG}C_{\\rm BG}=en_{\\rm B}+\\frac{\\mu_{\\rm B}}{e}(C_{\\rm BG}+C_{\\rm IL})-\\frac{\\mu_{\\rm T}}{e}C_{\\rm IL}$ and $V_{\\rm TG}C_{\\rm TG}=en_{\\rm T}+\\frac{\\mu_{\\rm T}}{e}(C_{\\rm TG}+C_{\\rm IL})-\\frac{\\mu_{\\rm B}}{e}C_{\\rm IL}$, where $C_{\\rm IL}$ is the interlayer capacitance between the two TBG subsystems. Because each layer's chemical potential appears in the other layer's gate equation only through $C_{\\rm IL}$, tracing a constant-$\\mu$ resistance line while the other layer is incompressible converts gate voltages into $(\\mu_{\\rm B}, n_{\\rm B})$ points, and the dimensions of incompressible-state diamonds in the $(V_{\\rm TG}, V_{\\rm BG})$ plane give the gap sizes $E_{g,\\nu}$. The extracted value $C_{\\rm IL}=1.6\\,\\mu{\\rm F/cm^2}$ corresponds to a 0.55 nm vacuum separation between the TBG mid-planes, supporting the weak-coupling picture.","core_discovery":"Two sets of flat bands, one from each twisted bilayer pair, coexist in a four-layer graphene double moiré system and can be gated separately. The paper's central claim is that because the middle interface has a large rotational mismatch, electron tunneling across it is suppressed and the two subsystems conserve their carrier numbers separately, interacting only electrostatically; this allows a thermodynamic analysis (Eq. 1) that converts resistance maps in the two-gate plane into chemical potential versus density curves for each TBG. The resulting data, spanning twist angles 0.91°–1.57°, show that correlated insulators at quarter-multiple moiré fillings appear only near the magic angle, whereas charge-neutrality gaps grow with twist angle away from it, and the hole-side flat band is flatter than the electron side. These trends are reproduced by a self-consistent Hartree calculation that uses a twist-angle-dependent ratio of same-sublattice to different-sublattice interlayer tunneling and a non-local tunneling term. No superconductivity was observed down to 95 mK in these samples, which the authors tie to the flatter valence band and to the different dielectric environment of the double-moiré stack.","pith_inferences":["A structural measurement (for example scanning tunneling or transmission electron imaging) could test the model's implicit assumption directly by checking whether lattice corrugation across the stack varies with twist angle in the way the tuned tunneling ratio $\\alpha$ implies.","If the middle twist angle were deliberately reduced below about 5°, the capacitance-only analysis should break down; the deviation would give a quantitative measure of residual interlayer tunneling between the two TBGs.","The same architecture could be reused with one TBG replaced by another kind of two-dimensional system, turning the remaining TBG into a built-in low-temperature chemical-potential sensor for its neighbor."],"forward_implications":["A single four-layer device can host two flat-band systems at different twist angles, so flat-band correlations can be studied in one stack without mixing the bands.","The extracted $\\mu$ versus $n$ curves across 0.91°–1.57° provide a thermodynamic reference for TBG flat bands away from the magic angle that agrees with earlier near-magic-angle measurements.","The different twist-angle dependence of charge-neutrality gaps and integer-filling gaps indicates that the two types of gaps come from different broken-symmetry states.","Because the valence flat band is flatter than the conduction band, the data support the view that superconductivity in these systems is favored on the valence side.","Correlated insulators survive in one TBG while another TBG sits nearby, showing that the proximity of a second flat-band system does not by itself destroy the correlated state."],"supporting_citations":[{"why":"Establishes the flat-band model of twisted bilayer graphene that the double-moiré system realizes twice.","marker":"[1]"},{"why":"Demonstrated the correlated insulating state at magic angle that this paper seeks in each constituent TBG.","marker":"[2]"},{"why":"Reported superconductivity in magic-angle twisted bilayer graphene, providing the reference behavior that the double-moiré stacks are checked against.","marker":"[3]"},{"why":"Shows how a double-layer system with separately conserved carrier densities can be used to probe chemical potentials electrostatically.","marker":"[25]"},{"why":"Provides a thermodynamic measurement of chemical potential versus filling in twisted bilayer graphene that the present method builds on.","marker":"[27]"},{"why":"Reports the cascade of chemical-potential jumps at fractional filling in magic-angle twisted bilayer graphene, the signature the paper reproduces in each subsystem.","marker":"[29]"},{"why":"Measures the chemical potential and bandwidth near the magic angle by compressibility, giving the comparison data for the extracted bandwidth changes.","marker":"[30]"},{"why":"Measures gap sizes at integer fillings in twisted bilayer graphene, the benchmark for the extracted $E_{g,\\nu}$ data.","marker":"[31]"},{"why":"Supplies the self-consistent Hartree calculation and the non-local interlayer tunneling model used for the theoretical comparison.","marker":"[40]"},{"why":"Introduces the non-local interlayer tunneling term that produces the electron-hole asymmetry in the calculated band structure.","marker":"[44]"}],"fun_headline_variants":["Four-layer graphene stack yields two independent flat bands","Double moiré hosts tunable flat bands per bilayer pair","Two flat bands, one stack: independent tuning in graphene","Graphene double moiré: separate flat bands, tunable each","Independent flat bands in a four-layer graphene stack"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Carrier number is separately conserved in each twisted bilayer pair, so the two systems interact only through an interlayer capacitance; if electrons tunnel across the middle interface in non-negligible amounts, the extracted chemical potentials and gap sizes shift.","fun_headline_variants_meta":{"raw":{"variants":["Four-layer graphene stack yields two independent flat bands","Double moiré hosts tunable flat bands per bilayer pair","Two flat bands, one stack: independent tuning in graphene","Graphene double moiré: separate flat bands, tunable each","Independent flat bands in a four-layer graphene stack"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000146,"raw_usage":{"total_tokens":1160,"prompt_tokens":903,"completion_tokens":257,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":175}},"tokens_in":519,"tokens_out":257,"duration_ms":3415,"temperature":1.0,"reasoning_tokens":175,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:53:46.988243+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be a double-moiré device with the middle twist angle deliberately reduced below about 5°: if interlayer tunneling matters, constant-$\\mu$ traces from the top layer would no longer give identical $\\mu_{\\rm B}$ versus $n_{\\rm B}$ curves at different top-layer fillings, and the diamond-edge slopes would deviate from the capacitance-only prediction. Alternatively, Landau-fan spectroscopy at matched densities could reveal tunnel-induced anticrossings between the two subsystems.","supporting_citations":[{"cited_title":"Kim , author I","cited_arxiv_id":null,"evidence_quote":"Shows how a double-layer system with separately conserved carrier densities can be used to probe chemical potentials electrostatically."},{"cited_title":"Weak Coupling Theory of Magic-Angle Twisted Bilayer Graphene","cited_arxiv_id":"2401.02872","evidence_quote":"Supplies the self-consistent Hartree calculation and the non-local interlayer tunneling model used for the theoretical comparison."}],"review_version":1}