{"id":"3f384eb4-6502-4f2f-a5fa-09ce31d36a63","arxiv_id":"2504.21651","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A one-shot GW calculation predicts that correlations remove the gamma hole pocket of La3Ni2O7 and that the La-5d x2-y2 band approaches the Fermi level under pressure and strain, matching ARPES without a Hubbard U.","lead":"Using a parameter-free GW calculation, this paper maps how pressure and strain reshape the Fermi surface of the superconducting nickelate La3Ni2O7. It argues that electron correlations, not a fitted Hubbard U, remove a contested hole pocket and that a lanthanum band moves low enough to matter at superconducting pressures.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fermi-level method ambiguity decides the claimed no-gamma Fermi surface at 29.5 GPa: Appendix D shows the DOS-integration E_F is 87 meV lower and reintroduces the gamma pocket, inside the paper's stated 100 meV uncertainty.","rationale":"The reader's weakest assumption identifies the same load-bearing concern, and I find it decisive. The paper is honest, detailed, and the ambient-pressure ARPES agreement plus the prior G0W0 result of Christiansson et al. are genuine supporting evidence. However, the 29.5 GPa Fermi-surface topology is exactly the quantity that changes when a different standard E_F method is used, and the paper provides no argument that selects the bisection Ts=10^-4 result over the DOS-integration result. The 30 meV indirect gap, the 70-80 meV GW shift, and the 48 meV E_F method spread are all inside the quoted 100 meV accuracy, so the no-gamma conclusion is not yet established at the claimed level. The 14 GPa La-5d crossing is explicitly hedged by the authors and is not needed for the main topological claim. I therefore keep the reader's conditional verdict: the calculation is plausible and well documented, but the central claim requires either a more robust Fermi-level determination or a demonstration that the topology is insensitive to the E_F method.","tokens_in":22584,"tokens_out":8080,"duration_ms":80683,"concrete_test":"Recompute the 29.5 GPa G0W0 Fermi level on the raw 6x6x6 GW eigenvalues with the tetrahedron method (zero smearing, electron-number conservation) and re-plot the Wannier-interpolated Fermi surface at that E_F. If the tetrahedron E_F falls at or below the DOS-integration value (48 meV below the Ts=10^-4 value), the gamma pocket reappears and the central no-gamma claim fails; if it falls inside the 30 meV indirect gap and reproduces Fig. 1b, the ambiguity is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that at 29.5 GPa the G0W0 Fermi surface has no gamma hole pocket and only alpha+beta sheets (Fig. 1b)—is settled by one Fermi-level choice. Section III A 2 reports an indirect gap of only 30 meV between the lower Ni-3d_z2 band top and the La-5d_x2-y2 band bottom, and the authors quote a 100 meV GW accuracy. Appendix D shows that the DOS-integration method places E_F 87 meV below the Ts=10^-2 bisection value and 48 meV below the Ts=10^-4 value adopted for Fig. 1b; that shift removes the lambda electron pocket and reintroduces the gamma hole pocket, restoring a DFT-like Fermi surface. The 48 meV difference is comparable to the 70-80 meV GW downward shift of the gamma band (Sec. III A 1) and sits at the edge of the stated uncertainty stripe in Fig. 1d. The paper is transparent about this ambiguity, but transparency does not make the topology robust: a different standard method for E_F gives qualitatively different low-energy physics. Since the ARPES agreement and the claim that no Hubbard U is needed both rest on the no-gamma topology, this Fermi-level sensitivity is the load-bearing weak point. The internal-coordinate sensitivity of Appendix B compounds the problem, but the E_F ambiguity alone is sufficient to make the central claim conditional.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents one-shot G0W0 calculations of the electronic structure of La3Ni2O7 at 29.5 GPa, at ambient and other pressures, and under biaxial strain. The central claim is that GW quasiparticle corrections shift the lower Ni-3d_z2 band down by about 70–80 meV, removing the gamma hole pocket from the Fermi surface without introducing any Hubbard U, so that the Fermi surface consists only of the cuprate-like beta sheet and the nickelate-specific alpha cylinder. A second, related claim is that the La-5d_{x^2-y^2} band is pulled down by correlations and pressure, approaching or crossing the Ni-3d_z2 bands near the pressure onset of superconductivity, leading to self-doping. The authors compare their results with ARPES data and with previous DFT, hybrid-functional, DMFT, and GW studies, and provide detailed appendices on structural sensitivity, Fermi-level determination, orbital character projections, and convergence.","tokens_in":22849,"tokens_out":2642,"duration_ms":29422,"significance":"If the central Fermi-surface topology is robust, the paper would be an important contribution: it would show that a parameter-free one-shot G0W0 calculation already captures the experimentally observed absence of the gamma pocket, and it would identify the La-5d_{x^2-y^2} band as a low-energy player relevant for pairing models. The work has real strengths: no fitted Hubbard U or hybrid mixing parameter enters the self-energy; the authors report convergence tests for the number of bands (Appendix G); they provide a detailed and transparent comparison of Fermi-level determination methods (Appendix D); and they carefully discuss the strong sensitivity to internal atomic coordinates (Appendix B). The comparison with ARPES is a useful benchmark. However, the central no-gamma topology is not shown to be robust with respect to the method used to compute the Fermi level, and the paper itself documents that a different standard choice changes the topology qualitatively. That issue is load-bearing for the main conclusion, so the paper needs additional work before the central claim can be accepted as established.","major_comments":[{"comment":"The claim that at 29.5 GPa the G0W0 Fermi surface has no gamma pocket and no lambda pocket depends on the choice of the Fermi level from the bisection method at T_s = 10^-4 Ha. Appendix D shows that the DOS-integration method places E_F 87 meV lower than the T_s = 10^-2 bisection value and 48 meV lower than the adopted T_s = 10^-4 value, which reintroduces the gamma hole pocket and removes the lambda electron pocket. Because the indirect gap is only 30 meV and the paper states a 100 meV GW accuracy, the no-gamma result is not robust to this standard methodological choice. I ask the authors to quantify the occupancy-consistent E_F and its uncertainty more rigorously, to show the Fermi surfaces obtained with each E_F method with the same smearing, and to state clearly which of their conclusions depend only on E_F-insensitive features such as the robust alpha and beta sheets.","section":"Appendix D and Figs. 1b, 1d"},{"comment":"The pressure-evolution plot in Fig. 4b assigns the crossing of the La-5d_{x^2-y^2} band with the Ni-3d_z2 bands to 14 GPa, but for the 14 and 40 GPa calculations the Fermi level was not actually computed; it was estimated by a linear interpolation from calculations where E_F was available, as stated in the figure caption. Since the pressure onset of superconductivity at 14 GPa is one of the paper's notable coincidences, this estimate needs an explicit uncertainty and ideally a direct E_F calculation for those pressures, because the 30 meV indirect gap at 29.5 GPa already lies below the stated GW accuracy.","section":"Section III A 2 and Fig. 4b"},{"comment":"The central Fermi-surface results rely on the choice of PBE-relaxed internal atomic coordinates for the 29.5 GPa structure, while Appendix B shows that using experimental XRD internal positions produces a fake La-5d cylindrical Fermi sheet at both DFT and GW levels. This is not merely a minor structural detail: the no-gamma, no-lambda topology at 29.5 GPa is defined relative to the relaxed coordinates, and the physical coordinates are not known to the required precision. The authors should state more explicitly how this structural uncertainty limits the low-energy predictions, and whether any conclusion about self-doping or the 14 GPa crossing survives across the range of plausible internal-position choices.","section":"Appendix B and Section II B"}],"minor_comments":[{"comment":"The sentence describing a 'shift of 0.1 eV to avoid poles/divergences' is not fully specified; please state whether this shift is applied to the Green's function or to the self-energy evaluation and at which frequencies.","section":"Section II A"},{"comment":"The caption says that dot width is proportional to orbital contribution, but in the printed figure many dots are difficult to distinguish; a supplementary color version or a zoomed inset around the Fermi level would improve readability.","section":"Fig. 1d caption"},{"comment":"The Fermi-surface panels in Fig. 10 are labeled only by the method names in the header row; please add explicit labels to each panel indicating which E_F value was used, since the difference between the T_s = 10^-4 and T_s = 10^-2 bisection cases is central to the discussion.","section":"Appendix D, Fig. 10"},{"comment":"The statement that the 6x6x6 k-sampling is not precise enough to resolve possible Fermi arcs is useful, but it would be helpful to quantify the k-point spacing in energy units near the Fermi surface so readers can judge the limitation.","section":"Section III A 4"}],"recommendation":"major_revision","confidential_remarks":"I see no problematic citation pattern or novelty concern. The main issue is internal: the paper's own Appendix D documents that the central Fermi-surface topology is not stable under a standard change of the Fermi-level determination method. The authors are transparent about this, but transparency does not remove the need to either establish robustness or substantially soften the central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The main thing to know: this is a transparent and well-documented G0W0 study of La3Ni2O7, but its headline result—that correlations remove the gamma pocket without a Hubbard U—is not robust at the pressures where superconductivity actually occurs. The paper itself shows this in Appendix D, and I think it deserves to be said plainly rather than buried.\n\nWhat is genuinely good: the calculation is full plane-wave G0W0, done at several pressures and for epitaxial strain, with convergence checks, careful orbital projections, and no fitted parameters. The La-5d x2-y2 band lowering is real and interesting, especially the coincidence with the 14 GPa onset of superconductivity. The negative result—that you don't need a large Hubbard U to kill the gamma pocket—is physically meaningful, even if the proof is softer than the abstract suggests. The paper is also admirably honest about its own fragility; that is not nothing.\n\nWhere it gets soft: at 29.5 GPa the claimed Fermi surface (no gamma, no lambda) depends on choosing the Fermi level from a bisection calculation at Ts = 10^-4 Ha. The alternative DOS-integration method, also standard, puts E_F 87 meV lower and brings back the gamma hole pocket while deleting the lambda electron pocket. The indirect gap is only 30 meV, and the paper's own stated accuracy is 100 meV. So the central topology is undetermined within the method's own uncertainty. The internal-coordinate dependence (Appendix B) compounds this: experimental XRD positions produce a fake La-5d cylinder, and the main text relies on PBE-relaxed coordinates. Neither choice is indefensible, but together they mean the central claim is conditional, not resolved. There is also a minor table-versus-text inconsistency in the 29.5 GPa lattice parameter that should be cleaned up.\n\nNovelty is moderate: Christiansson et al. already got gamma-pocket removal in a reduced G0W0 model, and You et al. noticed the La-5d shift. The new contribution is the full plane-wave treatment, the pressure/strain dependence, and the framing of the La-5d band as a low-energy actor.\n\nWho is this for? Anyone working on bilayer nickelate modeling. It belongs in the refereed literature, but the authors should be pushed to reframe the conclusion from \"GW resolves the gamma question\" to \"within GW accuracy the topology is unresolved, and the La-5d band is a candidate low-energy ingredient.\" That is an honest and still valuable message.\n\nMy recommendation: send it to review. A serious referee can push on the Fermi-level sensitivity and the structural assumptions without rejecting the work, and the transparency here is a useful model for the field.","headline":"Honest, carefully documented G0W0 study whose central no-gamma Fermi-surface claim at 29.5 GPa hinges on a Fermi-level method choice the paper itself shows to be ambiguous.","tokens_in":23448,"tokens_out":2194,"would_cite":false,"duration_ms":24706,"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":"The paper argues that GW correlations, without any Hubbard U, remove the gamma hole pocket from the Fermi surface of La3Ni2O7 and pull a La-5d band down to self-dope near 14 GPa.","keywords":["La3Ni2O7","bilayer nickelate superconductivity","GW approximation","Fermi surface topology","gamma hole pocket","self-doping","high pressure","epitaxial strain"],"falsifier":"An ARPES experiment on bulk La3Ni2O7 under 29.5 GPa that resolves the Brillouin-zone corner would settle the main claim: observing a gamma hole pocket at the Fermi level would contradict the central Fermi-surface topology. A complementary check is neutron diffraction that confirms the experimental XRD oxygen positions, which would make the predicted La-5d Fermi surface an artifact.","tokens_in":22379,"feed_emoji":"","tokens_out":5467,"duration_ms":53520,"temperature":0.7,"pith_summary":"The paper claims that the contested gamma hole pocket in La3Ni2O7 disappears once electronic correlations are included through the GW approximation, without any adjustable Hubbard U. It also claims that pressure and epitaxial strain pull down a La-5d x2-y2 band from the spacer layers so that it nearly reaches the Fermi level, crossing the Ni-3dz2 bands around 14 GPa, the pressure where superconductivity begins. A sympathetic reader would care because this would resolve a long-standing disagreement between DFT calculations and ARPES and point to a previously neglected low-energy band relevant for the pairing mechanism. The result is presented for the bulk at 29.5 GPa, as a function of pressure, and for a strained thin-film structure.","feed_headline":"GW calculation removes the gamma Fermi pocket in La3Ni2O7","feed_subtitle":"One-shot many-body calculation matches ARPES and finds a La-5d band self-doping near 14 GPa, the superconducting onset.","key_machinery":"The central object is the one-shot GW approximation (G0W0) to the many-body self-energy, applied on top of PBE-DFT with a Godby-Needs plasmon-pole model and Wannier interpolation of 67 projected Wannier functions. The G0W0 quasiparticle energies, not Kohn-Sham eigenvalues, are treated as the physical excitation energies comparable to ARPES. The argument depends on the difference between DFT and GW Fermi surfaces: correlations lower the Ni-3dz2 band and the La-5dx2-y2 band by different amounts, changing the Fermi-surface topology. The paper also uses band-projected orbital characters and quasiparticle spectral weight Z to identify the La-5d band and measure correlation strength, with Z dropping to about 0.67 on Ni-3d states.","core_discovery":"At the one-shot G0W0 level, correlation effects shift the lower Ni-3dz2 band down by about 70 to 80 meV along the X-P direction, enough to remove the gamma hole pocket from the Fermi level. The resulting Fermi surface has only the cuprate-shaped beta sheet and the nickelate-specific alpha cylinder, both of Ni eg character and effectively 1D, in agreement with ARPES. The same calculation lowers the La-5dx2-y2 band of the spacer atoms by about 0.7 eV at the M point, leaving an indirect gap of only 30 meV between the Ni-3dz2 top of valence and the La-5dx2-y2 bottom of conduction at 29.5 GPa. Because correlations and pressure act in the same direction, this La band crosses the Ni-3dz2 bands near 14 GPa and would begin to occupy states on La atoms, an effective self-doping of the NiO2 planes. The paper concludes that the ab initio GW electronic structure already matches experiment without Hubbard U or strongly correlated physics.","pith_inferences":["As an editorial inference, if the gamma pocket is truly absent, theoretical models that anchor pairing on the dz2 hole pocket may need to be re-examined; the paper itself does not make that claim.","The 14 GPa crossing of the La-5d band is suggestive but could be coincidence; a targeted extension would be to include this band explicitly in spin-fluctuation pairing calculations.","Because the Fermi topology changes radically with tiny internal atomic displacements, a neutron-diffraction determination of the oxygen positions could distinguish the predicted Fermi surface from the artifact-laden XRD structure.","A testable extension would be to track the predicted self-doping onset with pressure using optical or Hall measurements, since the paper identifies the La band occupation as the mechanism but does not compute transport signatures."],"forward_implications":["If the paper is correct, the gamma hole pocket does not exist at the Fermi level in bulk La3Ni2O7 at 29.5 GPa, so pairing models that rely on that pocket need revision.","The La-5dx2-y2 band is a low-energy degree of freedom that effective Ni-eg models omit; it becomes occupied above about 30 GPa, self-doping the NiO2 planes.","At 14 GPa, where superconductivity appears, the La-5dx2-y2 band crosses the upper Ni-3dz2 band, making the spacer band part of the low-energy physics at the onset.","The epitaxially strained thin-film case is electronically similar to bulk at roughly 14 GPa, so in-plane strain and pressure tune the same low-energy features.","The interlayer dz2 gap that enters some proposed pairing scenarios is left essentially unchanged by GW in this calculation, so that scenario survives.","The ideal no-gamma, no-lambda Fermi surface is a narrow window; defects or small doping shifts could open either pocket in a real sample."],"supporting_citations":[{"why":"Supplies the discovery of superconductivity in bulk La3Ni2O7, the experimental pressure range, and the XRD structure used in the comparison.","marker":"[1]"},{"why":"Provides the 29.5 GPa lattice parameters and internal atomic positions used in the benchmark calculation, as well as the dz2 interlayer pairing scenario that the paper compares against.","marker":"[24]"},{"why":"A previous G0W0 calculation on a reduced Ni-eg model that also found no gamma pocket, but whose EDMFT step reintroduced it; the paper positions its full-band result against this.","marker":"[45]"},{"why":"A recent G0W0 calculation that found an upward shift of the upper Ni-3dz2 band and highlighted the La-5d band, defining a discrepancy the paper addresses.","marker":"[15]"},{"why":"DFT and DMFT calculations that retain the gamma pocket while CDMFT removes it, supporting the paper's argument that nonlocal correlations are essential.","marker":"[55]"},{"why":"The ARPES experiment that finds no gamma pocket, providing the main experimental benchmark for the GW Fermi surface.","marker":"[56]"},{"why":"A DFT study linking the lowering of the La-5dx2-y2 band with pressure and the shape of the superconducting region, which the paper extends with GW.","marker":"[70]"},{"why":"Reports superconductivity under epitaxial compressive strain and provides the in-plane lattice parameter used in the thin-film simulation.","marker":"[5]"},{"why":"An ARPES experiment on strained films that reports the gamma pocket, used as a contrasting case for the strained-structure comparison.","marker":"[57]"}],"fun_headline_variants":["GW erases gamma pocket in La3Ni2O7","No Hubbard U: GW nails La3Ni2O7 Fermi surface","GW predicts La self-doping in La3Ni2O7 at 14 GPa","GW wipes gamma pocket, matches ARPES in La3Ni2O7","La3Ni2O7: GW gets Fermi surface without Hubbard U"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Fermi level at 29.5 GPa is correctly fixed by the bisection method at T_s = $10^{-4}$ Ha; the alternative DOS-integration method puts it 87 meV lower and reintroduces the gamma hole pocket, and the indirect gap of 30 meV is below the stated 100 meV uncertainty.","fun_headline_variants_meta":{"raw":{"variants":["GW erases gamma pocket in La3Ni2O7","No Hubbard U: GW nails La3Ni2O7 Fermi surface","GW predicts La self-doping in La3Ni2O7 at 14 GPa","GW wipes gamma pocket, matches ARPES in La3Ni2O7","La3Ni2O7: GW gets Fermi surface without Hubbard U"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001018,"raw_usage":{"total_tokens":4346,"prompt_tokens":1044,"completion_tokens":3302,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":660,"completion_tokens_details":{"reasoning_tokens":3201}},"tokens_in":660,"tokens_out":3302,"duration_ms":21965,"temperature":1.0,"reasoning_tokens":3201,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:56:53.892189+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An ARPES experiment on bulk La3Ni2O7 under 29.5 GPa that resolves the Brillouin-zone corner would settle the main claim: observing a gamma hole pocket at the Fermi level would contradict the central Fermi-surface topology. A complementary check is neutron diffraction that confirms the experimental XRD oxygen positions, which would make the predicted La-5d Fermi surface an artifact.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the 29.5 GPa lattice parameters and internal atomic positions used in the benchmark calculation, as well as the dz2 interlayer pairing scenario that the paper compares against."},{"cited_title":"We can expect that any tiny perturbation of the sys- tem, e.g","cited_arxiv_id":null,"evidence_quote":"A previous G0W0 calculation on a reduced Ni-eg model that also found no gamma pocket, but whose EDMFT step reintroduced it; the paper positions its full-band result against this."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The ARPES experiment that finds no gamma pocket, providing the main experimental benchmark for the GW Fermi surface."},{"cited_title":"coordinates can strongly affect the resulting electronic structure","cited_arxiv_id":null,"evidence_quote":"Reports superconductivity under epitaxial compressive strain and provides the in-plane lattice parameter used in the thin-film simulation."}],"review_version":1}