{"id":"0635a85c-77fb-4ba7-b4ec-8e15884214f7","arxiv_id":"2412.16316","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In particle-hole conjugated fractional quantum Hall states, interference and noise measurements indicate coherent tunneling of quasiparticle bunches with charge νe, which a local gate can dissociate back to elementary charges.","lead":"In interference experiments on fractional quantum Hall states at filling factors 2/3, 3/5, and 4/7, the authors observe flux periods and shot-noise Fano factors that imply quasiparticles tunnel as coherent bunches of two, three, or four elementary charges. Charging a small top gate in the interferometer dissociates these bunches and restores the usual elementary-charge period, an effect no current theory explains.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The dissociation observation is not yet distinguished from a top-gate-induced reduction of the effective interference area; the fixed-area flux conversion is the load-bearing assumption.","rationale":"I read the paper in good faith. The experimental observations are novel and reproducible across two devices, and the public data availability is genuine support. The Fano factor F=ν at VTG=0 is consistent with the bunched AB period and with previous low-temperature noise measurements, so the bunching claim has independent support. My concern is not about data quality or honesty. It is that the dissociation part of the central claim depends on the fixed-area flux conversion, and the internal modulation-gate data fail the same consistency check in two of three fillings. This is exactly where the weakest assumption sits. The manuscript flags the anomaly rather than hiding it, which is why the concern warrants a condition rather than a rejection. I agree with the reader's conditional verdict, but I would sharpen the condition: before acceptance, the authors should provide an in-situ area/frequency calibration (e.g., using the coexisting residual peaks) and reconcile the modulation-gate periodicities at ν=3/5 and 4/7 with the claimed charge switch.","tokens_in":13351,"tokens_out":11982,"duration_ms":110631,"concrete_test":"Re-analyze the raw B–VMG FFT data in the mixed regime where bunched and dissociated peaks coexist (ν=2/3 and 3/5 in Fig. 4 and Supp. S9; OMZI2 in Supp. S11b), using a joint two-peak fit within each dataset. Compute R_B = ΔB_bunched/ΔB_dissociated and R_V = ΔVMG_bunched/ΔVMG_dissociated. With fixed A=3 µm² and fixed capacitance, the charge-change model predicts R_B = R_V = q (q=3 at ν=3/5; q=4 at 4/7), while an area-change-only model predicts R_B = q and R_V = 1. Earlier data imply R_V ≈ 1.7 at ν=3/5, matching neither; if confirmed, the paper must supply a model that simultaneously explains R_B, R_V, and the Fano factor before the dissociation can be attributed to a change in interfering charge.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that every magnetic-field period be converted to a flux period through the fixed geometric area A=3 µm², and that charging the top gate changes the interfering charge but not the enclosed loop area. This fixed-area assumption is the least secure link. When the top gate is charged, the bunched period ΔΦ=Φ0/ν is replaced by ΔΦ=(e/e*)Φ0; the ratio is q=ν/(e*/e), equal to 2, 3, and 4 at ν=2/3, 3/5, and 4/7. A top gate that shrinks the effective loop area by the same factor q would produce exactly the same period change with the quasiparticle charge remaining νe. The top gate sits at the center of the interferometer and is deliberately used to form an antidot/dot, and for ν=4/7, A/q=0.75 µm² is close to the top-gate area (≈0.79 µm² in OMZI1), so this is a concrete alternative, not a formal one. The measured Fano factor, which stays F=ν after charging, is consistent with that alternative and does not select the charge-change interpretation unless one assumes unmeasured neutral-mode noise. The modulation-gate periodicity is the diagnostic that would distinguish area from charge, but it is anomalous: at ν=3/5 and 4/7, the dissociated ΔVMG values (17.4 mV and 15.5 mV) disagree with the fixed-area charge-change predictions (≈30.9 mV and ≈38 mV), a discrepancy the authors state is not understood. The dissociation claim therefore rests on an unverified conversion that the internal data do not independently confirm.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports Aharonov-Bohm (AB) interference and shot-noise measurements in a chiral Mach-Zehnder interferometer at the fractional quantum Hall fillings ν=2/3, 3/5, and 4/7. With the top gate uncharged, the authors observe magnetic flux periodicities ΔΦ=Φ0/ν and single-QPC Fano factors F=ν, which they interpret as coherent tunneling and interference of bunched quasiparticles with charge νe. When the top gate is charged, the AB periodicity changes to ΔΦ=(e/e*)Φ0, which they interpret as dissociation into elementary quasiparticles, while the Fano factor remains F=ν. The paper states that these two observations are not expected by current theories and discusses possible implications for other fractional quantum Hall states.","tokens_in":13566,"tokens_out":14028,"duration_ms":116844,"significance":"If the interpretation holds, the results would establish that the interfering quasiparticle charge in a fractional quantum Hall interferometer can be set by the bulk filling rather than by the elementary fractional charge, and that a local gate can switch between the two regimes. The manuscript provides public data, uses two devices, and includes useful controls such as different edge-mode configurations (Supp. S7) and a second smaller top-gate device (Supp. S11). However, the load-bearing conversion from magnetic-field periods to flux periods rests on a fixed loop area that is not independently verified in the dissociated regime, and the Fano-factor behavior in that regime is not fully consistent with the proposed picture. The strength of the evidence is therefore currently insufficient for the strongest claims made in the title and abstract.","major_comments":[{"comment":"The central dissociation claim depends on converting every magnetic-field period into a flux period using the fixed geometric area A=3 µm². If charging the top gate shrinks the effective loop area by the factor q=ν/(e*/e), the same period change would be observed with the quasiparticle charge unchanged; for ν=4/7 the required area of roughly 0.75 µm² is close to the top-gate area, making this a concrete alternative. The modulation-gate data provide an internal check, but they contradict the fixed-area charge-change prediction: in the dissociated regime the observed ΔVMG values at ν=3/5 and 4/7 are 17.4 mV and 15.5 mV, whereas scaling the bunched-regime values (10.3 mV and 9.5 mV) by q=3 and 4 predicts approximately 30.9 mV and 38 mV. The authors state this discrepancy is not understood. As written, the dissociation evidence is therefore not distinguished from a gate-induced area change, and the conclusion that the AB periodicity changes because the quasiparticle charge changes is not independently supported.","section":"Paragraph beginning 'Next, we charged the top gate...' and the paragraph beginning 'The modulation gate voltage…"},{"comment":"The Fano factor in the dissociated regime remains F=ν while the AB periodicity corresponds to the elementary charge e*, and the paper appeals to neutral modes to reconcile this. The manuscript itself notes that intermediate transmission plateaus were not observed in the OMZI's QPCs, so the neutral-mode contribution was not established. The sentence 'the agreement between the Fano factor and AB periodicity through the relation ΔΦ≈Φ0/F at VTG=0 strongly suggests an incoherent origin of the increased Fano factor (due to neutral modes)' is internally inconsistent: if the Fano factor had an incoherent neutral-mode origin, it would not determine the AB periodicity. The authors should clarify this argument and provide additional evidence, such as the VTG dependence of F measured in the full interferometer, before concluding that the tunneling charge changes while the Fano factor does not.","section":"Paragraph beginning 'The present experiment focused on the low-temperature regime...' and Supp. S6"},{"comment":"The interpretation of the data as coherent bunched quasiparticles of charge νe assumes that the AB phase is accumulated solely by the charge of the interfering quasiparticle, with no contribution from anyonic statistical phases or from bulk quasiparticles beneath the top gate. The paper notes that localized quasiparticles are expected to produce phase slips, yet none were observed in these particle-hole conjugated states. This absence, together with the unexplained modulation-gate behavior, leaves the proposed bunching/dissociation picture without a quantitative theoretical account. The authors should either provide a more detailed model for the observed periodicities or explicitly frame the result as an empirical observation that challenges current theories, rather than as a demonstrated property of bunched anyons.","section":"Paragraph beginning 'Our observations present a new paradigm...' and the Conclusions"}],"minor_comments":[{"comment":"The notation for the flux quantum is inconsistent: the abstract and figure captions use Φ0, while the main text frequently uses F0. Please unify the notation.","section":"Throughout"},{"comment":"The derivation of the modulation-gate periodicity is difficult to follow; the relation ΔVMG×nb×B=Φ0×P should be defined more carefully, with every symbol (including the parameter P) introduced and its physical meaning stated.","section":"Paragraph following Fig. 3"},{"comment":"The term 'pajama' is used for the AB interference pattern without definition; a brief explanation would make the paper more accessible to readers outside the immediate group.","section":"Fig. 1 and main text"},{"comment":"In the sentence 'the top gates radia are 0.5 µm (OMZI1) and 0.35µm (OMZI2)', 'radia' should be 'radii'.","section":"Methods / device parameters"},{"comment":"The note that 'the noise of the entire interferometer does not follow a simple relationship to the QP charge' is an important caveat and should be stated in the main text rather than only in the supplement.","section":"Supp. S6"}],"recommendation":"major_revision","confidential_remarks":"This is an interesting experimental paper with public data and several thoughtful controls, but the dissociation claim is not yet supported because of the unverified fixed-area conversion and the anomalous modulation-gate periods in the dissociated regime. The paper's reliance on its own prior Ref. 16 for the OMZI methodology is acceptable, but the novelty relative to that work should be clarified. I would recommend major revision, with the expectation that the authors can provide additional controls, clarify the Fano-factor argument, or soften the claims to match the strength of the evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a genuinely new experimental observation, not a rehash. Prior shot-noise-only works already measured F=ν at these fillings; what's new is the AB flux period Φ0/ν with the top gate uncharged and the gate-controlled crossover back to (e*/e)Φ0. The agreement between period and Fano factor at V_TG=0, plus the edge-mode control, makes the bunched-charge interpretation there reasonably credible. Data are public, two devices, real value.\n\nThe dissociation claim is where I get less comfortable. Charging the top gate changes the period to (e/e*)Φ0, and the ratio is exactly the factor q=ν/(e*/e) (2, 3, 4). A top gate that shrinks the effective interference area by the same factor q would produce the same period change while the interfering charge stays νe. The top gate sits at the center of the interferometer and is used to form an antidot/dot, so a change in effective loop area is not a contrived alternative. The modulation-gate periodicity is the diagnostic that would separate area from charge, but in the dissociated regime at ν=3/5 and 4/7 it disagrees with the fixed-area charge-change prediction, and the paper says this is not understood. That is a load-bearing soft spot, not a cosmetic one.\n\nAlso, in the dissociated regime the Fano factor remains F=ν while the AB period says e*. The authors explain this via neutral modes, but the neutral-mode contribution was not measured in this setup. They say so themselves. The relation ΔΦ=Φ0/F, which works at V_TG=0, is violated after charging, so the two measurements no longer point to the same charge.\n\nTo be fair: the paper flags all of this. It does not hide the missing pieces, and the fixed-area assumption is explicit. The gate-controlled crossover in Fig. 5 is interesting, and the one-by-one dissociation is a nice check. I just would not sign on to the coherent-bunched-anyons conclusion as established.\n\nWho should read it: anyone working on FQH interferometry or quasiparticle charge. It deserves a serious referee; the experimental result is likely real, and the interpretation has holes that referees can push on. My own verdict would be conditional: publish after the area alternative is addressed, or with the interpretation softened. I'd cite the data, but not the strong claim.","headline":"A likely real experimental effect in FQH interferometry whose interpretation as charge-bunched anyons is not yet nailed down; the dissociation claim rests on an unverified fixed-area conversion.","tokens_in":14289,"tokens_out":3420,"would_cite":true,"duration_ms":29036,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["73.43.-f","72.70.+m","73.23.-b"],"model":"deepseek-v4-flash","headline":"At filling factors 2/3, 3/5, and 4/7, a chiral interferometer measures flux periods set by the bulk filling ν, implying that bunched quasiparticles, not elementary ones, carry the interference.","keywords":["fractional quantum Hall effect","Aharonov-Bohm interference","anyons","quasiparticle bunching","Fano factor","Mach-Zehnder interferometer","particle-hole conjugate states","top-gate dissociation"],"falsifier":"Measure the same three fillings with a second interferometer of a different known area, or with an in-situ tunable area: if the periods are truly flux periods $\\Phi_0/\\nu$ and $(e/e^*)\\Phi_0$, the magnetic-field period must scale inversely with the enclosed area; if it does not, the fixed-area conversion is the source of the apparent effect. A second, complementary check is to sweep temperature through the range where earlier work saw $F$ change from $\\nu$ to $e^*/e$: the bunched periodicity should disappear in the same temperature window.","tokens_in":13025,"feed_emoji":"⚛️","tokens_out":12335,"duration_ms":96988,"temperature":0.7,"pith_summary":"This paper reports that in a chiral Mach-Zehnder interferometer at the fractional quantum Hall fillings $\\nu=2/3$, $3/5$, and $4/7$, the Aharonov-Bohm flux period is $\\Delta\\Phi=\\Phi_0/\\nu$ rather than $\\Delta\\Phi=(e/e^*)\\Phi_0$, and the shot-noise Fano factor at a single point contact (the inferred charge of the partitioned particles) is $F=\\nu$ rather than $F=e^*/e$. Because independent elementary quasiparticles would give both quantities in terms of the elementary charge $e^*$, the authors conclude that coherent bunches of two, three, or four elementary quasiparticles, carrying charge $e_b=\\nu e$ each, are the objects that tunnel and interfere. Charging a small top gate in the interferometer bulk dissociates the bunches and restores the elementary-charge flux period, while the Fano factor stays at $\\nu$, which the authors attribute to neutral edge modes adding incoherent noise. The two observations, bunching and dissociation, are not expected by current theory, and the authors suggest the same physics may appear in other fractional quantum Hall states at lower temperatures.","feed_headline":"Interference sees anyons tunneling as pairs, triples, quadruples","feed_subtitle":"Flux period and shot noise point to charge νe until a top gate restores the elementary-charge period.","key_machinery":"The load-bearing object is the bunched quasiparticle, defined operationally by the identity $\\Delta\\Phi=\\Phi_0/F$: the flux period and the Fano factor are consistent only with a tunneling charge $e_b=\\nu e$, i.e., a Laughlin quasiparticle comprising two, three, or four elementary anyons. The device that exposes it is the optical Mach-Zehnder interferometer (OMZI), whose two point contacts partition co-propagating edge modes so that all incoming charge must pass through the interference loop, producing clean Aharonov-Bohm 'pajama' patterns. The top gate is the control knob: left uncharged it leaves the bunches intact, and charged it forms a dot or antidot whose local quasiparticles dissociate the bunches and change the flux period without changing the point-contact noise. Neutral edge modes serve as the auxiliary mechanism invoked to keep $F=\\nu$ in the dissociated regime.","core_discovery":"The paper's central claim is that the objects that tunnel and interfere in these particle-hole-conjugated states are Laughlin quasiparticles of charge $e_b=\\nu e$, not the elementary quasiparticles of charge $e^*=e/3$, $e/5$, or $e/7$. The evidence is the joint pattern $\\Delta\\Phi=\\Phi_0/\\nu$ and $F=\\nu$ with the top gate uncharged, which satisfies $\\Delta\\Phi=\\Phi_0/F$ exactly as if a single object of charge $\\nu e$ carried the current. When the top gate is charged to form a dot or antidot, the flux period jumps to $\\Delta\\Phi=(e/e^*)\\Phi_0$, the value fixed by the elementary quasiparticles, while $F$ stays at $\\nu$; the authors read this as dissociation of the bunches, with neutral modes supplying the extra noise so that the Fano factor no longer tracks the interfering charge. The dissociation is not all-or-nothing: at intermediate top-gate voltages the two periodicities coexist, and at $\\nu=3/5$ the system evolves continuously from bunched to dissociated and back as the gate voltage is swept.","pith_inferences":["Beyond the paper: if bunching is generic, previous interferometric determinations of fractional charge that operated near the same temperatures may have actually measured $\\nu e$ rather than $e^*$; re-examining those datasets for the $\\Phi_0/\\nu$ period would test this.","Beyond the paper: the dissociation threshold should depend on the top gate's area and on the density of localized quasiparticles it creates; a systematic sweep of dot size and gate voltage could distinguish a number-fluctuation mechanism from a direct electrostatic effect on the bunches.","Beyond the paper: a testable prediction is that lengthening the interferometer arms, or equilibrating the edges, should change the neutral-mode contribution and therefore reduce $F$ toward $e^*/e$ in the dissociated regime while leaving the flux period unchanged."],"forward_implications":["With no top-gate charge, the relationship $\\Delta\\Phi=\\Phi_0/F$ identifies the interfering charge as $\\nu e$, so AB periodicity alone cannot be used to read off the elementary charge $e^*$ in these states at low temperature.","A single local gate can switch the interferometer between a bunched regime and a dissociated regime, giving experimental control over which quasiparticle species sets the interference period.","In the dissociated regime the Fano factor no longer tracks the interfering charge, showing that shot noise and AB periodicity can disagree when neutral modes contribute, a combination that any complete theory of these edges must reproduce.","The same low-temperature bunching should appear in Jain particle states and in even-denominator fractional quantum Hall states, with correspondingly modified flux periods.","Braiding experiments in non-Abelian states will need to account for the fact that the coherently interfering object can be a Laughlin quasiparticle, which is Abelian, even when elementary quasiparticles of the state are not."],"supporting_citations":[{"why":"Supplies the OMZI design and the particle-state baseline ($\\Delta\\Phi=(e/e^*)\\Phi_0$) that the new bunching and dissociation results are compared against.","marker":"[16]"},{"why":"Establishes the low-temperature Fano factor $F=\\nu$ in these fractional states and links it to the appearance of neutral modes.","marker":"[17]"},{"why":"Provides the stochastic-equilibration model used to argue that neutral-mode noise can make the Fano factor disagree with the tunneling charge.","marker":"[30]"},{"why":"Earlier shot-noise measurement at $\\nu=2/3$ that found the same Fano factor and is cited as agreement for the single-QPC noise data.","marker":"[33]"},{"why":"Prior evidence of scattering of bunched fractionally charged quasiparticles, which motivates reading $F=\\nu$ as bunching rather than elementary charge.","marker":"[20]"},{"why":"Recent bilayer-graphene interferometer reporting missing phase slips in particle-hole-conjugated states, cited as a possible analogue of the debunching mechanism.","marker":"[37]"},{"why":"Provides the integer quantum Hall electron-pair/triplet interference that the paper compares with and distinguishes from anyon bunching.","marker":"[12]"}],"fun_headline_variants":["Anyons bunch as νe pairs, triples, quads in interference","Top gate splits bunched anyons, restoring e* period","Bunched anyons dissociate to reveal elementary charge","Flux period and noise prove anyons tunnel in bunches","Surprise: anyons interfere as charge νe until a gate acts"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis converts every measured magnetic-field period into a flux period using a fixed enclosed area of $3\\,\\mu\\mathrm{m}^2$, and assumes the Aharonov-Bohm phase is accumulated solely by the tunneling charge with no contribution from anyonic statistical phase or from quasiparticles under the top gate.","fun_headline_variants_meta":{"raw":{"variants":["Anyons bunch as νe pairs, triples, quads in interference","Top gate splits bunched anyons, restoring e* period","Bunched anyons dissociate to reveal elementary charge","Flux period and noise prove anyons tunnel in bunches","Surprise: anyons interfere as charge νe until a gate acts"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000199,"raw_usage":{"total_tokens":1473,"prompt_tokens":1147,"completion_tokens":326,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":763,"completion_tokens_details":{"reasoning_tokens":237}},"tokens_in":763,"tokens_out":326,"duration_ms":3251,"temperature":1.0,"reasoning_tokens":237,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T10:43:10.650226+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same three fillings with a second interferometer of a different known area, or with an in-situ tunable area: if the periods are truly flux periods $\\Phi_0/\\nu$ and $(e/e^*)\\Phi_0$, the magnetic-field period must scale inversely with the enclosed area; if it does not, the fixed-area conversion is the source of the apparent effect. A second, complementary check is to sweep temperature through the range where earlier work saw $F$ change from $\\nu$ to $e^*/e$: the bunched periodicity should disappear in the same temperature window.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior evidence of scattering of bunched fractionally charged quasiparticles, which motivates reading $F=\\nu$ as bunching rather than elementary charge."}],"review_version":1}