{"id":"5a7bfe33-5e16-4154-a0c0-04c7af80a34f","arxiv_id":"2607.06492","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":4,"one_line_summary":"Charge transfer between TDBG and CrOCl enhances valley-polarized correlated insulating states at half-filling under high magnetic fields, without magnetic proximity effects.","lead":"This paper reports that placing twisted double bilayer graphene on a CrOCl substrate enhances correlated insulating states at half-filling via interfacial charge transfer rather than magnetic proximity. A generalist might read it because it shows a new way to tune electronic states in moiré materials using interface engineering rather than magnetic effects.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Displacement field confounds the charge-transfer attribution: the key comparison uses D=0.26 vs D=0.5 V/nm, and TDBG correlated states are strongly D-dependent.","rationale":"The reader correctly identified the most load-bearing concern: the confounding of displacement field with charge transfer in the key comparison. The paper compares D=0.26 V/nm (with CT) against D=0.5 V/nm (without CT), and since D is the dominant tuning knob for TDBG correlated states, this confound is real and significant. The reader also correctly noted the post-hoc g-factor fitting window issue and the authors' own acknowledgment that further experiments are needed. The CONDITIONAL verdict is appropriate: the experimental observation (enhanced resistance at half-filling in the CT region, reproducibility across cooldowns and sweep rates) is genuine and represents a real data point, but the mechanistic claim that charge transfer specifically causes the enhancement is not cleanly established given the D confound. I agree with the reader's assessment and see no need to change the verdict. The concern is not that the paper is wrong, but that the causal attribution is under-controlled. A TDBG/hBN control at matched D would settle whether the enhancement is due to the CrOCl interface (charge transfer) or to the displacement field itself.","tokens_in":10315,"tokens_out":1963,"duration_ms":138959,"concrete_test":"Fabricate a TDBG/hBN control device (no CrOCl) with the same twist angle (~1.1°) and measure the half-filling insulating state at D=0.26 V/nm under the same magnetic field range (0–25 T). If the control shows comparable resistance enhancement and g-factor at 20 T, the enhancement is a displacement-field effect, not charge transfer. If the control shows no enhancement at D=0.26 V/nm, the charge-transfer attribution is supported. Alternatively, if a second TDBG/CrOCl device can be fabricated where the charge-transfer boundary falls at a different D value, comparing across devices at matched D would also disentangle the two effects.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that charge transfer from TDBG to CrOCl enhances the half-filling correlated insulating state. The evidence rests on comparing the insulating state at ~20 T and the g-factor (18.1 vs 11.8) between two regions of the dual-gate map: D=0.26 V/nm ('with charge transfer') and D=0.5 V/nm ('without charge transfer'). However, these two regions differ not only in whether charge transfer occurs but also in displacement field, which is the primary tuning parameter for TDBG band structure and correlated states (refs 31, 32 in the paper). The displacement field changes the flatness of the moiré bands, the gap sizes, and the nature of correlated phases. The paper attempts to mitigate this by noting that 'the features of both diagrams on the hole-doped side are quite similar,' but hole-side and electron-side responses to D are generally asymmetric in TDBG due to band-structure asymmetry, so similarity on one side does not control for the other. Additionally, the g-factor comparison depends on post-hoc fitting windows ('within the range of magnetic field where the gap is developing and the valley polarization is dominant') that are not specified quantitatively for each case, making the 18.1 vs 11.8 comparison sensitive to analysis choices. The authors themselves acknowledge: 'To find out the underlying mechanism behind the enhanced insulating state in our samples, further experiments are needed.' The observation of enhanced resistance at 20 T in the CT region is real and reproducible (Appendix A), but the causal attribution to charge transfer rather than to the different displacement field is not cleanly isolated.","agreement_with_reader":"agree"},"referee_report":{"model":"glm-5.2","summary":"This manuscript reports transport measurements on twisted double bilayer graphene (TDBG, twist angle ~1.1°) fabricated on the antiferromagnetic insulator CrOCl. The authors observe a charge-transfer (CT) boundary in the dual-gate resistance map, separating regions where electrons are injected into CrOCl from those where they are not. The central experimental finding is an enhanced correlated insulating state at half-filling that reemerges at ultrahigh magnetic fields (~20 T) in the CT region, accompanied by a larger extracted g-factor (18.1 vs. 11.8) compared to the non-CT region. Reproducibility is demonstrated across different gate-sweep rates and cooling cycles (Figs. 5, 6). The authors attribute the enhancement to a superlattice Coulomb potential from charge transfer, while acknowledging that further experiments are needed to confirm the mechanism.","tokens_in":10533,"tokens_out":2264,"duration_ms":192295,"significance":"The study of interfacial charge-transfer effects on correlated states in moiré systems is a timely and relatively unexplored direction. The observation of a reentrant half-filling insulating state at ~20 T in TDBG/CrOCl is a genuine experimental result, and the reproducibility checks (Appendix A, Figs. 5–6) across sweep rates and a four-month-separated cooling cycle strengthen the claim that the feature is robust. The in-plane field suppression data (Fig. 4) provide a useful additional characterization. The work is a reasonable contribution to the emerging interface-engineering approach to moiré correlated states, though the causal attribution (see below) limits the strength of the conclusions at present.","major_comments":[{"comment":"§II, Figs. 2(a)–(b): The central claim that charge transfer enhances the half-filling insulating state rests on comparing D = 0.26 V/nm ('with CT') and D = 0.5 V/nm ('without CT'). These two points differ in displacement field, which is itself the primary tuning parameter for TDBG band structure and correlated states (the paper cites Refs. 31, 32 for this). The authors note that 'the features of both diagrams on the hole-doped side are quite similar,' but TDBG is known to exhibit strong electron–hole asymmetry in its D-dependent band structure, so similarity on the hole side does not control for D-dependent effects on the electron side where the enhancement is observed. This confound is load-bearing for the attribution of the enhanced insulating state to charge transfer rather than to the different displacement field. The authors should either (i) identify a comparison at the same (or相近)","section":null},{"comment":"§II, Fig. 3(b) and surrounding text: The g-factor comparison (18.1 vs. 11.8) depends on fitting windows described only qualitatively as 'the range of magnetic field where the gap is developing and the valley polarization is dominant.' The specific field ranges used for each linear fit are not stated, and Fig. 3(b) appears to show non-monotonic gap behavior (rising to ~20 T then decreasing). The extracted g-factors are thus sensitive to the chosen window. The authors should specify the exact field ranges used for each fit, show the linear fits overlaid on the data, and provide error bars or sensitivity analysis (e.g., how the g-factor changes if the window is shifted by ±1–2 T). Without this, the 18.1 vs. 11.8 comparison—and the conclusion that CT enhances valley polarization—is not independently verifiable.","section":null},{"comment":"§II, paragraph on mechanism: The paper offers two possible mechanisms (superlattice Coulomb potential from CT, and magnetic proximity hybridization) but does not present any evidence that would distinguish them or even favor one. Given that no magnetic-proximity-related phenomenon is observed (as the authors state), and the CT effect is the dominant observed interfacial effect, the attribution to the superlattice Coulomb potential is presented as the leading candidate but without direct evidence (e.g., dependence on CT strength, comparison across twist angles, or theoretical modeling of the expected gap enhancement). The authors' own statement that 'further experiments are needed' is appropriate, but the manuscript title and abstract ('Enhanced by Interfacial Effect') present a stronger causal claim than the data support. The authors should either temper the causal language or provide a控","section":null}],"minor_comments":[{"comment":"Appendix B, Eqs. for n and D: The definitions use n_0 and D_0 as offsets but do not state how these offsets were determined (e.g., from the CNP position at zero field). Please specify.","section":null},{"comment":"Fig. 1(c): The red dashed CT boundary is described qualitatively. It would help to state the approximate carrier density or gate-voltage coordinates at which this boundary occurs.","section":null},{"comment":"Fig. 1(d): The Chern insulator with C = 2 is mentioned and shown in Fig. 1(e), but the Hall plateau quantization in Fig. 1(e) is not clearly resolved to the precision typically expected for a Chern number extraction. Please comment on the plateau quality and how C = 2 was determined.","section":null},{"comment":"§II, Fig. 4: The quarter-filling state is mentioned but not discussed in detail. A brief comment on its nature (valley-polarized? spin-polarized?) and why it survives in-plane field while the half-filling state does not would strengthen the narrative.","section":null},{"comment":"Typographical: 'refridgerator' → 'refrigerator' (§II); 'enchanced' → 'enhanced' (Fig. 2 caption, Fig. 5 caption); 'comaparison' → 'comparison' (Appendix A); 'dahsed' → 'dashed' (Fig. 4 caption).","section":null},{"comment":"The twist angle is stated as 'around 1.1 degrees' extracted from Brown-Zak oscillations. The Supplementary Materials are referenced but not provided; please ensure the angle extraction and its uncertainty are documented there.","section":null}],"recommendation":"major_revision","confidential_remarks":"The displacement-field confound is the key issue. If the authors can find a within-device comparison that holds D approximately constant while varying CT (e.g., by scanning carrier density across the CT boundary at fixed D), this would substantially strengthen the paper. If such data are not available, the paper could still be acceptable with appropriately tempered claims (framing the observation as 'correlated with the CT region' rather than 'enhanced by charge transfer') and with the g-factor analysis made quantitative. The current title and abstract overstate what the data demonstrate."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful reading and constructive comments. The referee raises three major points: (1) the displacement-field confound in the CT vs. non-CT comparison, (2) the need for explicit g-factor fitting windows and sensitivity analysis, and (3) the causal language in the title and abstract. We agree that points (1) and (2) require revision and that point (3) warrants tempering of the causal language. We address each below.","responses":[{"response":"The referee correctly identifies that the comparison between D = 0.26 V/nm (with CT) and D = 0.5 V/nm (without CT) confounds the charge-transfer effect with the displacement-field dependence of the TDBG band structure. We acknowledge that the hole-side similarity does not rigorously control for D-dependent effects on the electron side, given the known electron–hole asymmetry in TDBG. Ideally, one would compare the same displacement field with and without CT; however, this is not possible within a single device because the CT boundary is fixed by the electrostatics of the TDBG–CrOCl interface. The CT region occupies a specific range of D, and the non-CT region necessarily lies at a different D. We will add an explicit discussion of this confound in the revised manuscript, acknowledging it as a limitation of the single-device comparison. We will also note that the most direct way to disentangle the two effects would be to compare against a TDBG-on-hBN control device at the same D, which we plan for future work. We agree this weakens the strength of the causal attribution and will adjust the language accordingly.","revision_made":"partial","referee_comment":"§II, Figs. 2(a)–(b): The central claim that charge transfer enhances the half-filling insulating state rests on comparing D = 0.26 V/nm ('with CT') and D = 0.5 V/nm ('without CT'). These two points differ in displacement field, which is itself the primary tuning parameter for TDBG band structure and correlated states. The authors note that 'the features of both diagrams on the hole-doped side are quite similar,' but TDBG is known to exhibit strong electron–hole asymmetry in its D-dependent band structure, so similarity on the hole side does not control for D-dependent effects on the electron side where the enhancement is observed. This confound is load-bearing for the attribution of the enhanced insulating state to charge transfer rather than to the different displacement field. The authors should either (i) identify a comparison at the same (or相近) displacement field with and without CT,"},{"response":"The referee is correct that the fitting windows for the g-factor extraction are not specified and that the non-monotonic behavior of the gap makes the result sensitive to the chosen range. We will revise the manuscript to include: (i) the exact magnetic field ranges used for each linear fit, (ii) the linear fits overlaid on the data in Fig. 3(b), and (iii) a sensitivity analysis showing how the extracted g-factors change when the fitting window is shifted by ±1–2 T. We agree that without this information the comparison is not independently verifiable. If the sensitivity analysis shows substantial variation in the g-factors, we will accordingly temper the quantitative claim and emphasize the qualitative trend (larger gap in the CT region) rather than the precise numerical values.","revision_made":"yes","referee_comment":"§II, Fig. 3(b) and surrounding text: The g-factor comparison (18.1 vs. 11.8) depends on fitting windows described only qualitatively as 'the range of magnetic field where the gap is developing and the valley polarization is dominant.' The specific field ranges used for each linear fit are not stated, and Fig. 3(b) appears to show non-monotonic gap behavior (rising to ~20 T then decreasing). The extracted g-factors are thus sensitive to the chosen window. The authors should specify the exact field ranges used for each fit, show the linear fits overlaid on the data, and provide error bars or sensitivity analysis (e.g., how the g-factor changes if the window is shifted by ±1–2 T). Without this, the 18.1 vs. 11.8 comparison—and the conclusion that CT enhances valley polarization—is not independently verifiable."},{"response":"We agree with the referee that the title and abstract present a causal claim stronger than what the data rigorously support. The data show a correlation between the CT region and the enhanced insulating state, but we cannot definitively attribute the enhancement to the superlattice Coulomb potential mechanism without additional evidence (e.g., dependence on CT strength, twist-angle comparison, or theoretical modeling). We will temper the causal language in the title and abstract. Specifically, we will change the title to use 'Associated with' or 'Correlated with' rather than 'Enhanced by,' and we will revise the abstract to state that the enhanced insulating state is observed in the CT region and that the superlattice Coulomb potential is a candidate mechanism, while explicitly noting that a direct causal link is not established. We will retain the discussion of both candidate mechanisms and the statement that further experiments are needed.","revision_made":"yes","referee_comment":"§II, paragraph on mechanism: The paper offers two possible mechanisms (superlattice Coulomb potential from CT, and magnetic proximity hybridization) but does not present any evidence that would distinguish them or even favor one. Given that no magnetic-proximity-related phenomenon is observed (as the authors state), and the CT effect is the dominant observed interfacial effect, the attribution to the superlattice Coulomb potential is presented as the leading candidate but without direct evidence (e.g., dependence on CT strength, comparison across twist angles, or theoretical modeling of the expected gap enhancement). The authors' own statement that 'further experiments are needed' is appropriate, but the manuscript title and abstract ('Enhanced by Interfacial Effect') present a stronger causal claim than the data support. The authors should either temper the causal language or provide a控"}],"tokens_in":10372,"tokens_out":1274,"duration_ms":168007,"standing_objections":[]},"desk_editor":{"model":"glm-5.2","letter":"Here's the short version: this paper reports the first TDBG/CrOCl heterostructure, showing a reentrant correlated insulating state at ~20 T and a g-factor enhancement (18.1 vs 11.8) in the charge-transfer region. The observation is real and reproducible. The causal attribution to charge transfer is confounded by the experimental design, and the authors half-admit it themselves. The paper is honest experimental work with an incompletely justified mechanistic claim. It deserves a serious referee because the data are worth engaging with, but the framing needs to be substantially tightened before publication. The stress-test concern about displacement field confounding is correct and is the central issue. The key comparison uses D=0.26 V/nm (with charge transfer) vs D=0.5 V/nm (without), and TDBG band structure and correlated states are strongly D-dependent (the paper cites Koshino and Chebrolu et al. on exactly this point). The authors' defense is that hole-side features look similar in both regions, but electron-hole asymmetry in TDBG means hole-side similarity doesn't control for the electron side where the enhanced insulating state appears. So the confound is real and not adequately addressed. The g-factor comparison (18.1 vs 11.8) inherits the same problem. Additionally, the fitting windows for the Arrhenius slopes are described qualitatively (where the gap is developing and valley polarization is dominant) but not specified quantitatively for each case, making the comparison sensitive to analysis choices. The authors acknowledge this by stating further experiments are needed to pin down the mechanism. What's genuinely new: extending the charge-transfer physics from bilayer graphene/CrOCl to TDBG, and the reentrant insulating state at ultrahigh fields in TDBG is a new data point. The reproducibility data in Appendix A (different sweep rates, different cooling cycles four months apart) is solid and gives confidence the observation is not an artifact. The in-plane field dependence showing suppression of the half-filling state is a nice additional result suggesting valley-spin coupling. This is for experimentalists working on moiré systems and interface engineering. The data are worth a referee's time. Recommend would_accept_peer_review=true, with the expectation that the displacement-field confound and g-factor fitting windows are addressed head-on, either with additional controls or by substantially softening the mechanistic claims.","headline":"Real observation of reentrant correlated insulating state in TDBG/CrOCl, but the charge-transfer attribution is confounded by displacement field","tokens_in":11180,"tokens_out":567,"would_cite":false,"duration_ms":148451,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["73.21.-b","73.22.-f","71.70.Di"],"model":"glm-5.2","headline":"Charge transfer, not magnetism, boosts correlated insulation in graphene","keywords":[],"falsifier":"If the enhanced insulating state and larger g-factor could be reproduced at D=0.26 V/nm in a TDBG device on a non-charge-transferring substrate (e.g., plain hBN), the attribution to charge transfer would be undermined. Conversely, if the enhancement appears only in the presence of a charge-transferring substrate and scales with the degree of charge transfer, the claim is strengthened.","tokens_in":10585,"feed_emoji":"⚡","tokens_out":1061,"duration_ms":146366,"temperature":0.7,"pith_summary":"This paper reports that when twisted double bilayer graphene (TDBG) is placed on the antiferromagnetic insulator CrOCl, the dominant interfacial effect is not magnetic exchange but charge transfer: electrons leak from the graphene into the CrOCl substrate, and the resulting interfacial charge order acts as a superlattice Coulomb potential on the graphene layers. The authors observe that in the charge-transfer region, the correlated insulating state at half-filling of the moiré band is strongly enhanced at ultrahigh magnetic fields around 20 T, with a thermal activation gap whose g-factor is 18.1 versus 11.8 in the region without charge transfer. They interpret the larger g-factor as evidence that charge transfer amplifies the valley polarization of the correlated state. The paper also reports a Chern insulator with C=2 emerging from the half-filling state, suppression of the half-filling state by in-plane magnetic field (suggesting valley-spin coupling), and reproducibility of the enhanced insulating state across different cooling cycles and gate-sweeping rates. The central claim is that interfacial charge transfer provides a route to manipulate correlated states in moiré graphene systems, distinct from and potentially more impactful than magnetic proximity effects.","feed_headline":"Charge transfer, not magnetism, boosts correlated insulation in graphene","feed_subtitle":"Twisted double bilayer graphene on CrOCl shows enhanced insulating states driven by interfacial electron transfer, offering a new knob for m","key_machinery":"The mechanism involves electrons transferring from TDBG into CrOCl, creating a long-wavelength electronic crystal (charge order) at the CrOCl surface. This charge order serves as a superlattice Coulomb potential on the graphene layers, enhancing bandgap opening at half-filling. The valley-polarized nature of the correlated state is diagnosed through the magnetic-field dependence of the thermal activation gap (Arrhenius fitting), yielding g-factors that quantify the coupling between the gap and the out-of-plane field. The comparison between charge-transfer and non-charge-transfer regions at different displacement fields (0.26 V/nm vs 0.5 V/nm) is the experimental lever used to isolate the two","core_discovery":"The paper's central finding is that charge transfer between TDBG and CrOCl enhances the valley-polarized correlated insulating state at half-filling, as demonstrated by a larger thermal activation gap (g=18.1 vs 11.8) and higher resistance at 20 T in the charge-transfer region compared to the region without charge transfer. The authors attribute this enhancement to a superlattice Coulomb potential generated by long-wavelength charge order at the CrOCl surface, which acts on the graphene layers and amplifies gap opening. Notably, no magnetic exchange effect from the antiferromagnetic substrate is detected; the interfacial charge-related effect dominates.","pith_inferences":[],"forward_implications":["If charge transfer is a generalizable knob for enhancing correlated states, other graphene/magnetic-insulator heterostructures previously studied for magnetic proximity effects may need re-examination through the lens of interfacial charge redistribution.","The superlattice Coulomb potential mechanism suggests that engineering the charge-transfer partner (substrate choice, interlayer spacing, twist angle relative to the substrate) could tune the strength and periodicity of the effective potential on the moiré bands.","The observed valley-spin coupling (suppression of the half-filling state by in-plane field) hints at a richer internal structure of the correlated state than pure valley polarization, which could be relevant for spin-valley entangled phases in moiré systems.","The reentrant insulating state at ultrahigh field (around 20 T, near half-flux per moiré unit cell) connects to the broader phenomenology of reentrant correlated insulators in twisted bilayer graphene, suggesting a shared mechanism across different moiré platforms."],"fun_headline_variants":["Interfacial charge transfer boosts correlated insulation in twisted graphene","CrOCl interface amplifies correlated insulating states in twisted graphene","Charge transfer, not magnetism, drives correlated states in graphene moiré","Graphene insulation boosted by interfacial charge transfer at CrOCl","Correlated insulation in twisted graphene enhanced by interfacial charge transfer"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The paper compares the charge-transfer region (displacement field D=0.26 V/nm) with the non-charge-transfer region (D=0.5 V/nm) and attributes the enhanced insulating state to charge transfer, but the two regions differ in displacement field, which independently modifies TDBG band structure and correlated states. Without independently controlling displacement field from the charge-transfer boundary, the causal role of charge transfer is not cleanly separated from the effect","fun_headline_variants_meta":{"raw":{"variants":["Interfacial charge transfer boosts correlated insulation in twisted graphene","CrOCl interface amplifies correlated insulating states in twisted graphene","Charge transfer, not magnetism, drives correlated states in graphene moiré","Graphene insulation boosted by interfacial charge transfer at CrOCl","Correlated insulation in twisted graphene enhanced by interfacial charge transfer","Charge transfer amplifies correlated insulation in graphene moiré systems"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":1254,"prompt_tokens":523,"completion_tokens":731,"prompt_tokens_details":null},"tokens_in":523,"tokens_out":731,"duration_ms":22258,"temperature":1.0,"reasoning_tokens":666,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-08T03:58:49.477287+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If the enhanced insulating state and larger g-factor could be reproduced at D=0.26 V/nm in a TDBG device on a non-charge-transferring substrate (e.g., plain hBN), the attribution to charge transfer would be undermined. Conversely, if the enhancement appears only in the presence of a charge-transferring substrate and scales with the degree of charge transfer, the claim is strengthened.","supporting_citations":[],"review_version":1}