{"id":"f0d55d4b-ed30-4087-a683-f0cd1e2e8ba1","arxiv_id":"1909.01501","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Circular polarization resolved magneto-infrared spectroscopy of multilayer epitaxial graphene reveals a four-fold splitting of the monolayer n=0 Landau level transition, with extracted valley and Zeeman g-factors of 6.7 and 4.8.","lead":"This paper uses circularly polarized infrared light and magnetic fields up to 17.5 T to study multilayer epitaxial graphene, and resolves four separate energy levels within one Landau level transition of its monolayer regions. The authors interpret these as spin and valley symmetry breaking and extract effective g-factors of 6.7 and 4.8.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The A–D modes are not independently shown to come from one monolayer layer; without that, the four-fold splitting and both g-factors are not established.","rationale":"The reader identified the same load-bearing assumption: the A–D modes are all assigned to sub-LL transitions of a single high-mobility monolayer graphene layer, and this is not independently verified. I agree with that assessment and refine it by noting that the paper's own evidence of sample heterogeneity (MLG, BLG, and the ** low-mobility mode) makes the multi-layer alternative concrete. The paper's key argument for Zeeman splitting is the claimed linear-in-B dependence of Δ<, which depends on cancellation of √B terms in Eq. (4); because the fitting parameters and uncertainties are not reported, this cancellation is not demonstrated. The same concern affects the extracted g-factors, since they are obtained from the same decomposition. This is a real but addressable problem: a quantitative refit of the published data, or a control experiment on a known single monolayer, could settle it. The paper is otherwise internally consistent and reports plausible data, so the conditional verdict is appropriate rather than accept or reject. No independent fatal flaw was identified.","tokens_in":11061,"tokens_out":8164,"duration_ms":83311,"concrete_test":"Digitize the A–D transition energies from Fig. 2(b); fit T_i = a_i√B + b_iB for each mode and compute (a_A − a_B + a_C − a_D)/2 with full covariance. If this √B coefficient is significantly nonzero, Δ< is not purely linear in B, and the two-layer Fermi-velocity alternative remains viable. Additionally, test whether the best-fit a_i values correspond to a common vF consistent with the reported veF ≈ 1.025×10^6 m/s and vhF ≈ 0.975×10^6 m/s; if the four modes require incompatible Fermi velocities, the single-layer sub-level assignment is not supported by the data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on assigning the four modes A–D (Fig. 1(d)) to four sub-Landau-level transitions of one high-mobility monolayer layer. Section 3.2 introduces this as 'one can attribute the A, B, C, and D modes to the four-fold splitting of the n = 0 transition in high mobility graphene layers,' while the ** mode is assigned to low-mobility layers. Because the same MEG sample contains MLG and BLG inclusions (Section 3.1) and at least one additional monolayer-like population (**), the alternative that A and B (electron-like) or C and D (hole-like) originate from different monolayer inclusions with different Fermi velocities or local doping is not excluded. In that case the energy separations would scale as δvF√(2eħB), and the claimed Zeeman splitting Δ<∝B would be an artifact of the flexible two-parameter fit T_i=a_i√B+b_iB in Eq. (5). The paper does not report the fitted a_i and b_i values, their uncertainties, or a test that the √B coefficients cancel in Eq. (4), so the linear-in-B identification of Δ< is not independently verified. Since g*ZS and g*VS are extracted from this same decomposition, the four-fold splitting and both g-factors fall if the single-layer assignment fails.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports circular-polarization-resolved magneto-infrared transmission measurements on multilayer epitaxial graphene (MEG) at 4.2 K in magnetic fields up to 17.5 T, using tunable quantum cascade lasers. For the monolayer graphene (MLG) inclusions, the n=0 to n=1 Landau-level transition is reported to resolve into four fine-structure modes, labeled A-D, attributed to the lifting of the valley and spin degeneracy of the zeroth Landau level together with electron-hole asymmetry. The transition energies are fit as T_i = a_i*sqrt(B) + b_i*B (Eq. 5), and the combinations Delta_< = [(T_A-T_B)+(T_C-T_D)]/2 and Delta_eh = [(T_A+T_B)-(T_C+T_D)]/2 are extracted (Eq. 4). A linear-in-B Delta_< is identified with an enhanced Zeeman splitting (g*_ZS = 4.8), while the larger splitting Delta_> is ascribed to a valley-symmetry-breaking state; the authors prefer a linear-in-B scenario (g*_VS = 6.7), while acknowledging that a sqrt(B) scenario fits equally well within their (unquantified) experimental uncertainty. For the bilayer graphene (BLG) inclusions, the paper reports electron-hole mass asymmetry of the opposite sign to exfoliated BLG, attributed to interlayer couplings in the MEG stack.","tokens_in":2291,"tokens_out":3105,"duration_ms":175786,"significance":"If established, the four-fold splitting of the zeroth-Landau-level transition in near-neutral monolayer graphene probed by bulk optical spectroscopy would be a significant result for the broken-symmetry phase diagram of graphene, complementing prior transport and scanning-probe studies. The manuscript has clear strengths: the algebraic extraction of Delta_< and Delta_eh from four transition energies is exact given the model of Eq. (3); the use of circular polarization to separate electron-like and hole-like transitions is appropriate; the internal consistency check on Delta_eh is reasonable; and the text is candid that the nature of Delta_> cannot be conclusively determined (Sec. 3.3). The reported effective g-factors are fit results rather than independent predictions, and the four-mode spectrum itself is a falsifiable observation that does not depend on the parameter values. The BLG sign anomaly is an interesting, independently testable claim. These strengths, however, are conditioned on the single-layer assignment of the A-D modes and on an error analysis that the manuscript does not currently provide.","major_comments":[{"comment":"The central claim rests on the assumption that modes A-D belong to a single high-mobility monolayer layer, but this is asserted ('one can attribute the A, B, C, and D modes to the four-fold splitting of the n = 0 transition in high mobility graphene layers') rather than demonstrated. The sample demonstrably contains multiple monolayer-like populations with different Fermi velocities: v_eF = 1.025x10^6 m/s, v_hF = 0.975x10^6 m/s (fitted in Sec. 3.3) and the ** mode at vF = 1.00x10^6 m/s. A 2.5% difference in vF between two layers would shift the n = 0 transition energy by about 3.5 meV at 14 T (where E is near 140 meV), the same order as the claimed splittings (Delta_< is about 2.2 meV and Delta_> about 5.4 meV at 14 T). The paper should therefore report the fitted a_i and b_i of Eq. (5) with their uncertainties and show explicitly that the sqrt(B) terms cancel in the combination Delta_< = [(T_A-T_B)+(T_C-T_D)]/2, i.e., that a_A - a_B + a_C - a_D is consistent with zero; otherwise the 'linear-in-B' identification of Delta_< and both quoted g-factors inherit an unconstrained systematic error. A quantitative argument excluding the possibility that A and B (or C and D) originate from different monolayer inclusions with slightly different vF or local doping is required.","section":"Sec. 3.2, Eqs. (4)-(5), Fig. 1(d)"},{"comment":"No error bars, standard deviations, or confidence intervals are given anywhere in the paper, yet the central conclusions are quantitative: Delta_< = 0.16 meV/T, Delta_> = 1.44 meV/sqrt(T) or 0.39 meV/T, and the derived g-factors. The B-field range over which the n = 0 transition is observed is only 11-17 T, where sqrt(B) and B are strongly correlated, so distinguishing Delta_< proportional to B from Delta_< proportional to sqrt(B) requires quantitative uncertainties. The text states that the two Delta_> scenarios cannot be differentiated 'within the experimental uncertainty' without quantifying that uncertainty, and the abstract nevertheless presents g*_VS = 6.7 (one of the two equally good fits) as the result; the abstract should carry the caveat stated in Sec. 3.3, where the authors write that 'this work cannot give a conclusive answer to this question.' In addition, the claimed consistency check that Delta_eh is proportional to sqrt(B) is partially tautological, since the fit ansatz of Eq. (5) guarantees a sqrt(B) component; the paper should verify that the sqrt(B) coefficient of Delta_eh equals (v_eF + v_hF)*sqrt(2ehbar) using the velocities independently fitted to the n = 1 and n = 2 transitions in Fig. 3.","section":"Sec. 3.3 and Fig. 2(c)"},{"comment":"The identification of the four modes A-D rests on correlating kinks with peaks in second derivatives of transmission spectra. No line-shape fits, no sample-to-sample reproducibility data, and no estimate of the peak-position uncertainty are provided, although subsequent analysis quotes transition energies to the precision needed for a 0.16 meV/T slope. The dismissal of the * mode as circular-polarization leakage is also unquantified; the degree of circular polarization of the setup is not reported. The authors should document how peak positions (and their uncertainties) were extracted, how many independent photon energies contributed to Fig. 2(b), and justify the leakage assignment quantitatively.","section":"Sec. 3.2, Fig. 1(d)"},{"comment":"The derivation of g*_ZS = 4.8 is not transparent as written: the text states that Delta_< = 0.16 meV/T corresponds to a g-factor of 2.8 and that adding the first Landau level's bare g-factor of 2 gives g*_ZS = 4.8. Since Eq. (3) neglects first-LL splittings, the measured Delta_< of Eq. (4) already contains both the zeroth-LL and first-LL Zeeman contributions; the sign convention connecting them (e.g., Delta_<,meas = |g_1 - g_0| mu_B B for spin-conserving transitions) must be specified to justify g_0 = 4.8 rather than g_0 = 0.8. This affects a headline number and should be clarified.","section":"Sec. 3.3, g*_ZS derivation"}],"minor_comments":[{"comment":"The typeset fractions are ambiguous; write Delta_< = [(T_A-T_B)+(T_C-T_D)]/2 and Delta_eh = [(T_A+T_B)-(T_C+T_D)]/2.","section":"Eq. (4)"},{"comment":"State how many independent photon energies and field points define Delta_< and Delta_eh, and add error bars or a bootstrap uncertainty band.","section":"Fig. 2(c)"},{"comment":"The estimate e^2/4pi*epsilon*epsilon_0*l_B approximately equal to 11 meV/sqrt(T) assumes epsilon approximately equal to 5; the provenance of this dielectric constant and the sensitivity of the estimate to it should be stated.","section":"Sec. 3.3"},{"comment":"The equivalence of sweeping the field in positive versus negative directions with sigma+ and sigma- illumination deserves an explicit sentence, since the equivalence holds only if the sample response is isotropic under field reversal.","section":"Sec. 3.1"},{"comment":"The BLG band masses m*_e = 0.0376m_0 and m*_h = 0.0283m_0 and the +/-14% asymmetry are quoted without uncertainties, despite being the basis for the sign-anomaly claim.","section":"Sec. 3.4"},{"comment":"The phrase 'best fit yields effective g-factors' overstates the determination of Delta_>; wording such as 'consistent with' would better reflect the caveat stated in Sec. 3.3.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for a solid-state physics journal and the experiment is competently executed. The abstract is somewhat more assertive than the body, which explicitly concedes that the two Delta_> scenarios cannot be distinguished; I would ask the editor to ensure the revised abstract reflects that caveat. The absence of any error analysis is the most fixable deficiency; the harder issue, proving that A-D come from one monolayer layer, may require additional data, but at minimum the authors should report the Eq. (5) fit parameters and the cancellation test described in Major Comment 1. I see no grounds for rejection, but the load-bearing assumptions need work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe genuinely new result here is that CP-resolved magneto-IR, on quasi-neutral multilayer epitaxial graphene, appears to resolve four sub-modes of the n=0 to n=1 Landau level transition in the monolayer inclusions, and the paper uses them to extract effective g-factors (g*ZS=4.8, g*VS=6.7). That would be a first for a bulk-sensitive optical measurement. The paper also reports a striking opposite-sign electron-hole asymmetry in the bilayer inclusions, which is a solid observation in its own right.\n\nThe analysis itself is clean in places. Equations (3) and (4) define Δ< and Δeh in terms of the four measured transition energies, and the algebra is straightforward. The authors are also appropriately hedged about their inability to distinguish Δ>∝√B from Δ>∝B, and they say the work cannot conclusively identify the symmetry-broken ground state. They give their best-fit parameters rather than hiding them.\n\nThe soft spot the stress-test identifies is real. The four modes A-D are assigned to four sub-LL transitions of one high-mobility monolayer layer, but the same MEG sample contains multiple monolayer-like populations and a broad ** mode attributed to lower-mobility layers. Nothing in the data independently shows that A and B (or C and D) come from the same layer rather than from two different monolayer inclusions with different Fermi velocities or local doping. If that were the case, the energy separations would scale as δvF√(2eħB), and the apparent linear-in-B Zeeman splitting would be an artifact of the flexible fit T_i=a_i√B+b_iB. The paper does not report the fitted a_i and b_i values or their uncertainties, and the B range (11-17 T) is narrow. The Δeh consistency check (√B behavior) is suggestive but not a test of the single-layer assignment.\n\nThat is the main issue. The g-factors and the four-fold splitting claim both fall if the assignment fails. So the paper's central conclusion is conditional, not proven. I don't think that makes it a weak paper; the experiment is difficult and the observation is likely real. But it needs a serious referee and a revision that quantifies uncertainties, reports the fit parameters, and ideally tests the single-layer interpretation (e.g., by checking the field behavior of the mode separations or by studying a sample with a known single monolayer region).\n\nVerdict: send to peer review, but the referee should press on the assignment.","headline":"A potentially important CP-resolved magneto-IR measurement of four-fold LL splitting in epitaxial graphene, but the interpretation hinges on an unverified single-layer assignment and the g-factors lack quantified uncertainty.","tokens_in":11946,"tokens_out":3280,"would_cite":false,"duration_ms":32610,"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":"This paper reports a four-fold splitting of the n=0 to n=1 Landau-level transition in monolayer inclusions of multilayer epitaxial graphene, attributed to the lifting of valley and spin degeneracy of the zeroth Landau level combined with…","keywords":["epitaxial graphene","Landau levels","valley splitting","Zeeman splitting","magneto-infrared spectroscopy","circular polarization","electron-hole asymmetry","quantum cascade laser"],"falsifier":"Take the four modes A-D at a series of photon energies and fit each one independently to $T_i = a_i\\sqrt{B} + b_iB$; if they all yield the same Fermi velocity and extrapolate to a common zero-field crossing consistent with one monolayer, the four-fold-splitting interpretation is supported, whereas distinct Fermi velocities or field scalings would show that the modes originate from different layers. A complementary check is scanning tunneling spectroscopy on the same multilayer sample at high field, which would image four resolved sub-LL peaks in a single monolayer rather than peaks spread across multiple layers.","tokens_in":10890,"feed_emoji":"🪲","tokens_out":7830,"duration_ms":71272,"temperature":0.7,"pith_summary":"The paper sets out to show that the zeroth Landau level of monolayer graphene is four-fold split in high magnetic fields and that this broken symmetry can be seen in bulk optical absorption, not only in transport. Using circular-polarization-resolved magneto-infrared spectroscopy on multilayer epitaxial graphene, the authors resolve four distinct absorption modes within the n=0 to n=1 Landau-level transition and attribute them to lifting of the valley and spin degeneracy of the zeroth Landau level together with electron-hole asymmetry. The magnetic-field dependence of the modes yields an enhanced Zeeman-like splitting and a larger valley-splitting contribution, with best-fit effective g-factors $g^*_{ZS}=4.8$ and $g^*_{VS}=6.7$. The bilayer inclusions of the same sample show electron-hole asymmetry of the opposite sign, which the authors trace to the stacking environment.","feed_headline":"Magneto-IR reveals four-fold splitting of graphene's n=0 Landau level","feed_subtitle":"Circularly polarized light separates valley and spin contributions, with effective g-factors of 6.7 and 4.8.","key_machinery":"The load-bearing mechanism is the circular-polarization selection rule for inter-Landau-level transitions: $\\sigma+$ light activates $LL_{-s}\\rightarrow LL_{s+1}$ and $\\sigma-$ light activates $LL_{-s-1}\\rightarrow LL_s$, so the two polarizations isolate electron-like and hole-like transitions and expose electron-hole asymmetry. The quantitative analysis centers on a four-sub-LL model of the zeroth Landau level with energies $\\pm(\\Delta_> \\pm \\Delta_<)/2$ and the two extraction formulas $\\Delta_< = [(T_A - T_B) + (T_C - T_D)]/2$ and $\\Delta_{eh} = [(T_A + T_B) - (T_C + T_D)]/2$, where $T_A\\ldots T_D$ are the measured transition fields interpolated by $T_i = a_i\\sqrt{B} + b_iB$. This decomposition separates the Zeeman-like linear-in-$B$ part from the valley-splitting part and tests whether the valley splitting scales as $\\sqrt{B}$ or $B$.","core_discovery":"On its own terms, the paper's central claim is that the monolayer inclusions of multilayer epitaxial graphene host a four-fold splitting of the n=0 Landau-level transition: four modes labeled A, B, C, and D appear in the magneto-transmission around the n=0 to n=1 transition, with two electron-like ($\\sigma+$) and two hole-like ($\\sigma-$) transitions. The energies are modeled with sub-LLs at $\\pm(\\Delta_> \\pm \\Delta_<)/2$, where $\\Delta_>$ and $\\Delta_<$ are the larger and smaller of the valley and Zeeman splittings of the zeroth LL, and the electron and hole first Landau levels are offset by different Fermi velocities, $v^e_F = 1.025\\times 10^6$ m/s and $v^h_F = 0.975\\times 10^6$ m/s. The data give $\\Delta_< = 0.16$ meV/T (linear in $B$), i.e. a Zeeman-like splitting with $g^*_{ZS}=4.8$ after including the bare electron g-factor of the first LL, and a remaining splitting $\\Delta_>$ that is consistent with either $\\sqrt{B}$ or $B$ dependence; under the linear-in-$B$ assumption the effective valley g-factor is $g^*_{VS}=6.7$, favoring an electron-phonon mechanism (Kekulé distortion or charge-density wave) over a spin-polarized ferromagnetic ground state.","pith_inferences":["A sharper test of the same-layer assumption would be to extract each mode's $\\sqrt{B}$ slope separately; identical Fermi velocities for A-D would support the four-fold splitting picture, while different slopes would indicate modes from distinct layers with different mobilities or densities.","Extending the same CP-resolved magneto-IR technique to higher photon energies or higher magnetic fields could settle whether $\\Delta_>$ scales as $\\sqrt{B}$ or $B$, which the current data cannot distinguish.","The reported effective g-factors are substantially larger than the bare electron value, so the implied interaction enhancement could be checked against transport measurements on the same MEG material in regimes where the Fermi level sits inside the split sub-LL gap.","The sign reversal of electron-hole asymmetry in bilayer inclusions suggests a design rule: the local rotation angle of neighboring layers may tune the sign and magnitude of band asymmetry in epitaxial graphene stacks."],"forward_implications":["If the assignment holds, bulk circular-polarization magneto-IR becomes a direct probe of broken-symmetry Landau-level states in graphene, complementing transport and tunneling measurements that may be complicated by edge effects.","The electron-hole asymmetry of the monolayer inclusions is quantified as about 2.5% in Fermi velocity, giving microscopic parameters for band-structure models of epitaxial graphene.","The linear-in-$B$ splitting $\\Delta_<$ rules out a spin-polarized ferromagnetic ground state of charge-neutral graphene in this system, consistent with prior transport and spectroscopy results.","A linear-in-$B$ valley splitting, if confirmed, would point to electron-phonon-driven symmetry breaking rather than a purely Coulomb exchange mechanism.","The opposite-sign electron-hole asymmetry in bilayer inclusions suggests that rotational stacking and neighboring layers can control the sign of band asymmetry, which could be used in band engineering."],"supporting_citations":[{"why":"Establishes the multilayer epitaxial graphene platform as quasi-neutral with record-high mobility, the sample type used here.","marker":"[26]"},{"why":"Provides the prior magneto-IR identification of the monolayer and AB-stacked bilayer Landau-level sequences in MEG that the spectral assignment builds on.","marker":"[42]"},{"why":"Earlier circular-polarization-resolved magneto-IR study of thin MEG whose opposite asymmetry pattern is contrasted to justify the new layer attribution.","marker":"[37]"},{"why":"Prior bulk magneto-optical study reporting a broken-symmetry state in MEG, framing the interpretation of the n=0 splitting.","marker":"[15]"},{"why":"Theory companion to [15] modeling symmetry-breaking states and valley splitting in MEG; supplies candidate ground states.","marker":"[16]"},{"why":"Theory paper classifying possible symmetry-breaking states of charge-neutral graphene and their magnetic-field dependences, used to assign $\\Delta_<$ and $\\Delta_>$.","marker":"[39]"},{"why":"Review source for the expected $\\sqrt{B}$ and $B$ scaling of Zeeman and valley splittings used in the interpolation.","marker":"[20]"},{"why":"Source for the monolayer Landau level energy expression (Eq. 1) that anchors the transition assignments.","marker":"[46]"},{"why":"Transport study favoring a canted antiferromagnetic state, which the present data's valley-splitting scenario contrasts with.","marker":"[13]"},{"why":"Tunneling evidence for a Kekulé-distortion ground state, one of the candidate mechanisms for the linear-in-$B$ valley splitting.","marker":"[19]"}],"fun_headline_variants":["Magneto-IR reveals g-factors 6.7 and 4.8 for valley and spin in graphene","Four-fold splitting of graphene n=0 Landau level from valley and Zeeman effects","Graphene's zero LL splits into four: valley and spin g-factors measured","Circular polarization resolves four-fold LL splitting: g*V=6.7, g*Z=4.8"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"All four modes A, B, C, and D come from the same high-mobility monolayer graphene layer and represent sub-Landau-level structure; if they actually come from different layers with different Fermi velocities or carrier densities, the four-fold splitting attribution fails.","fun_headline_variants_meta":{"raw":{"variants":["Magneto-IR reveals g-factors 6.7 and 4.8 for valley and spin in graphene","Four-fold splitting of graphene n=0 Landau level from valley and Zeeman effects","Graphene's zero LL splits into four: valley and spin g-factors measured","Circular polarization resolves four-fold LL splitting: g*V=6.7, g*Z=4.8"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000937,"raw_usage":{"total_tokens":4052,"prompt_tokens":1033,"completion_tokens":3019,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":649,"completion_tokens_details":{"reasoning_tokens":2916}},"tokens_in":649,"tokens_out":3019,"duration_ms":21325,"temperature":1.0,"reasoning_tokens":2916,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:15:48.204123+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the four modes A-D at a series of photon energies and fit each one independently to $T_i = a_i\\sqrt{B} + b_iB$; if they all yield the same Fermi velocity and extrapolate to a common zero-field crossing consistent with one monolayer, the four-fold-splitting interpretation is supported, whereas distinct Fermi velocities or field scalings would show that the modes originate from different layers. A complementary check is scanning tunneling spectroscopy on the same multilayer sample at high field, which would image four resolved sub-LL peaks in a single monolayer rather than peaks spread across multiple layers.","supporting_citations":[{"cited_title":"45, Physics of Solid Surfaces, Subvolume B-VIII, edited by G","cited_arxiv_id":null,"evidence_quote":"Establishes the multilayer epitaxial graphene platform as quasi-neutral with record-high mobility, the sample type used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the prior magneto-IR identification of the monolayer and AB-stacked bilayer Landau-level sequences in MEG that the spectral assignment builds on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier circular-polarization-resolved magneto-IR study of thin MEG whose opposite asymmetry pattern is contrasted to justify the new layer attribution."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior bulk magneto-optical study reporting a broken-symmetry state in MEG, framing the interpretation of the n=0 splitting."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Theory companion to [15] modeling symmetry-breaking states and valley splitting in MEG; supplies candidate ground states."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Theory paper classifying possible symmetry-breaking states of charge-neutral graphene and their magnetic-field dependences, used to assign $\\Delta_<$ and $\\Delta_>$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Review source for the expected $\\sqrt{B}$ and $B$ scaling of Zeeman and valley splittings used in the interpolation."},{"cited_title":"Orlita, C","cited_arxiv_id":null,"evidence_quote":"Source for the monolayer Landau level energy expression (Eq. 1) that anchors the transition assignments."},{"cited_title":"Zhang, Z","cited_arxiv_id":null,"evidence_quote":"Transport study favoring a canted antiferromagnetic state, which the present data's valley-splitting scenario contrasts with."},{"cited_title":"Orlita, L","cited_arxiv_id":null,"evidence_quote":"Tunneling evidence for a Kekulé-distortion ground state, one of the candidate mechanisms for the linear-in-$B$ valley splitting."}],"review_version":1}