{"id":"5a2dd2f1-37e7-4d1e-89a4-8e89e409ba51","arxiv_id":"2412.12972","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Using 3266 days of IceCube data, no dark matter annihilation signal from the Earth's center was found, and the resulting spin-independent scattering limits are the strongest among neutrino telescopes for WIMP masses above 100 GeV.","lead":"This paper reports a search for neutrinos from dark matter particles annihilating at the center of the Earth using ten years of IceCube data. No signal was found, and the resulting upper limits on how strongly dark matter interacts with ordinary matter are the strongest from any neutrino telescope for dark matter particles heavier than 100 GeV.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"MC-only background model for the fixed vertical source is not validated by an independent off-source control region; the quoted limits could shift if the simulated vertical background shape is wrong.","rationale":"The reader identified the MC-only background assumption as the weakest point, and I agree that this is the central structural assumption. The added specificity is that the most serious risk is not simply 'MC might be wrong' but that there is no independent, signal-free control region with identical local detector response. The final data/MC comparisons are performed after background normalizations are fit to the same data, so they cannot validate the shape at the signal direction. That said, I do not see evidence that the MC is in fact wrong, and the null result would likely survive modest background-shape changes. My concern is therefore strongest for the quantitative claim of world-leading limits: a shape bias of order 10-20% in the vertical background could move the limits outside the quoted comparison band. A control-region closure test, using the already-withheld verification data or an off-vertical zenith band, is a concrete way to settle whether this concern is realized. Because the paper is otherwise careful and the analysis procedure is standard, the appropriate response is to accept conditionally on this validation check rather than to reject or to demand a full reanalysis now.","tokens_in":18726,"tokens_out":9289,"duration_ms":104963,"concrete_test":"Use the withheld 353-day verification dataset, or an off-vertical zenith band (for example 160-170 degrees) as a control region, with the final LE and HE selections and all signal PDFs set to zero. Fit only the background nuisance parameters in Eq. (6) to this control region, freeze them, and then predict the observed event rate in the signal band (170-180 degrees). If the predicted rate agrees with data within the combined statistical and systematic uncertainties, the MC background shape is validated and the limits stand. If it does not agree, the limits on sigma_SI must be recomputed with a corrected vertical background model or with an explicit shape uncertainty on the signal-region bins.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The analysis is forced to rely on Monte Carlo for all background shapes because the source at the Earth's center is fixed in local coordinates and right-ascension scrambling cannot be used (Sec. 3). This is not merely a normalization uncertainty: the signal occupies a narrow region of the reconstructed-zenith/energy plane near vertical, while the background PDFs for that region come entirely from Corsika, Genie, and NuGen simulations (Sec. 4). The final data/MC comparisons in Figs. 6 and 7 are made after fitting the nuisance parameters in Eq. (6) to the same data. A fit with several normalization nuisance parameters can absorb a systematic shape error in the signal region while still looking acceptable globally. Unlike analyses using RA scrambling, there is no off-source sample with the same local detector response to validate the simulated vertical background. In particular, misreconstructed atmospheric muons and the high-zenith atmospheric neutrino spectrum are the least constrained components here. If the simulated background shape near 180 degrees is biased, the 90% C.L. limits on sigma_SI derived from Eq. (9) would shift, and the headline claim of world-leading neutrino-telescope limits for m_chi > 100 GeV could be affected. This is a structural limitation, not an observed discrepancy, but it is the most load-bearing assumption in the paper.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a search for muon neutrinos from WIMP annihilation at the center of the Earth using 3266 days of IceCube data spanning May 2011 to May 2020. Signal events are simulated with WimpSim/PYTHIA for three annihilation channels (τ+τ−, W+W−, bb̄) and masses from 10 GeV to 10 TeV; backgrounds (atmospheric muons, atmospheric neutrinos, astrophysical neutrinos) are modeled with Corsika, Genie, and NuGen. A dedicated two-tier event selection (low-energy and high-energy BDTs, with roughly 90-95% neutrino purity) is described in Sec. 4. The statistical analysis (Sec. 5) is a binned Poisson profile likelihood in reconstructed zenith and energy, with nested background fractions as profiled nuisance parameters (Eqs. 5-8), Gaussian constraints on background normalizations, and discrete systematic variations. No significant excess is found: the most significant result is χχ → bb̄ at mχ = 250 GeV with a post-trial significance of 1.06σ (Sec. 6). The paper derives 90% C.L. upper limits on the spin-independent WIMP-nucleon cross section (Fig. 8) and claims these are world-leading among neutrino telescopes for mχ > 100 GeV.","tokens_in":18993,"tokens_out":16900,"duration_ms":149345,"significance":"If the results hold, the paper delivers the strongest existing neutrino-telescope constraints on spin-independent WIMP-nucleon scattering for mχ > 100 GeV, improving on the earlier one-year IceCube search by up to an order of magnitude, and its limit formalism (Eq. 9 and the σSI conversion of Sec. 6) is recastable to other dark matter models. The statistical treatment is a genuine strength: a profile-likelihood framework with a nested background decomposition, pseudo-experiment-based sensitivity studies, trial-corrected significances, KDE smoothing with cross-validated bandwidths, Gaussian constraints on astrophysical and Genie/NuGen normalizations, and a broad set of discrete systematic variations covering ice properties, DOM efficiency, atmospheric models, and oscillation parameters. The withholding of a 353-day verification dataset is also good practice. The principal weakness is structural and is explicitly acknowledged in Sec. 3: because the source is fixed in local coordinates, right-ascension scrambling is impossible, and the background shape in the signal region is entirely Monte-Carlo-based; the data/MC agreement shown in Figs.","major_comments":[{"comment":"The background model in the signal region is entirely simulation-based, and the uncertainty attached to its shape is load-bearing for the headline limits. As stated in Sec. 3, because the source is fixed at the Earth's center, right-ascension scrambling cannot be used and 'we must rely on Monte Carlo simulations to model the background, for the optimization of the event selection, and for the statistical analysis.' The data/MC comparisons in Figs. 6 and 7 are made after the nuisance parameters η of Eq. (6) are profiled to the same data, so normalization-type nuisance parameters can absorb a systematic shape error in the near-vertical bins; the discrete variations among atmospheric models do not by themselves validate the least-constrained components, namely mis-reconstructed vertical muons and the high-zenith atmospheric neutrino spectrum. Since the 90% C.L. limits of Eq. (9) and the claimed world-leading sensitivity for mχ > 100 GeV follow directly from this background model, I request a quantitative robustness check: (i) a data/MC comparison or goodness-of-fit in the signal region evaluated with the nuisance parameters fixed at their null-hypothesis values, or using the withheld 353-day verification dataset described in Sec. 3; and (ii) an explicit evaluation of the shift in the 90% C.L. limits when an additional shape nuisance is introduced in the vertical region, for example a freely or loosely constrained normalization of the mis-reconstructed-muon component or a tilt of the background zenith distribution near θ ≈ 180°.","section":"Sec. 3; Sec. 5 (Eqs. 5-8); Figs. 6-7; Eq. (9)"}],"minor_comments":[{"comment":"The post-trial significance of 1.06σ is reported without any description of the trial-correction procedure; please specify the number of trials (masses, channels, and the LE/HE selection choice), the method (for instance pseudo-experiments), and how correlations among the trials are handled.","section":"Sec. 6"},{"comment":"The effective volume Veff appearing in Eq. (9) is never defined; please state how Veff is computed for each selection and WIMP mass and how it depends on the reconstructed-energy range and signal spectrum, since the flux limit is directly proportional to this quantity.","section":"Eq. (9)"},{"comment":"The claim that the limits are world-leading among neutrino telescopes for mχ > 100 GeV is substantiated in Fig. 8 only against ANTARES and the previous IceCube search; Super-Kamiokande (Ref. [23]) is cited in the introduction but is absent from the comparison, so the comparison should be completed or the claim narrowed.","section":"Fig. 8"},{"comment":"The sentence stating that 'the sum of the parameters, (ξ, η), is equal to 1' is imprecise: in the nested formulation of Eq. (6), the background fractions sum to (1 − ξ) rather than to 1; please clarify the normalization convention.","section":"Sec. 5, Eq. (6)"},{"comment":"The optimization of the final BDT score threshold is described only qualitatively; please state whether the threshold is chosen per mass/channel and whether the withheld verification sample was used in this optimization to avoid overfitting.","section":"Sec. 4"},{"comment":"The caption of Fig. 5 says 'The LE signal baseline and the atmospheric background are shown,' but the figure displays binned PDFs for the HE analysis; please correct the caption.","section":"Fig. 5 caption"}],"recommendation":"major_revision","confidential_remarks":"To the editor: this is a standard IceCube collaboration analysis that fits the journal scope. The reader recommended accept, but I recommend major revision because the single most load-bearing assumption of the analysis - the Monte-Carlo-based background shape in the near-vertical signal region - is not validated by any independent control sample, and its possible effect on the 90% C.L. limits is unquantified. The requested checks (a verification-sample cross-check and a shape-nuisance study) are feasible within the manuscript's scope. I see no issues with novelty disclosure or citation practice."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should know this is a straightforward extension paper, but a good one. The collaboration takes the one-year Earth-center WIMP search and pushes it to ten years (3266 days), adds reconstructed energy as an observable, and re-optimizes the event selection. The result is what they claim: for WIMP masses above 100 GeV, these are the best limits from any neutrino telescope, roughly an order of magnitude better than their own previous result and ANTARES. The most significant excess is 1.06 sigma post-trial, so it is a clean null.\n\nThe statistics are careful. Binned Poisson likelihood, profiled nuisance parameters, Gaussian constraints from external measurements, trial-corrected p-values, pseudo-experiment sensitivities, and a sensible split into low- and high-energy selections. Signal PDFs come from WimpSim/PYTHIA, backgrounds from Corsika/Genie/NuGen; the conversion from flux limit to sigma_SI is not circular. The systematics treatment, especially around ice optics, DOM efficiency, and neutrino generators, is thorough.\n\nThe soft spot is structural, and the paper does not hide it. Because the source is the center of the Earth, the signal sits at a fixed local direction (zenith ~180 degrees), so right-ascension scrambling cannot be used to build a background from an off-source region. Everything about the vertical background shape comes from Monte Carlo. The data/MC plots in Figs. 6 and 7 are made after the nuisance parameters have been fitted to the same data, so a shape error in the signal region could be partially absorbed and the plots would still look reasonable. Misreconstructed down-going muons and the high-zenith atmospheric neutrino spectrum are the least constrained pieces. If those simulated shapes are off, the limits would shift.\n\nHow much does this matter? For the central claim, only mildly. The limits are still the best among neutrino telescopes even with some shape uncertainty, and direct detection is far more sensitive for most of the mass range anyway. The paper is honest about the reliance on MC and does what it can with alternative atmospheric models and discrete systematics. I would not call this a flaw that undermines the result; it is a limitation that should be stated clearly in the paper, and it is.\n\nBottom line: this is a solid, useful contribution for anyone working on WIMP searches with neutrino telescopes or on capture/annihilation in the Earth. It deserves review and publication. If I were refereeing, I would ask for a little more discussion about how the fitted nuisance parameters might mask a vertical shape error, but I would not ask for new analysis.\n\nRecommendation: send to peer review.","headline":"A careful 10-year IceCube null search that delivers the strongest neutrino-telescope limits for heavy WIMPs; the MC-only background model is a real limitation, but honestly handled.","tokens_in":19498,"tokens_out":2360,"would_cite":true,"duration_ms":21728,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Ten years of IceCube data find no dark matter signal from Earth's center, and set the best neutrino-telescope limits on WIMP-nucleon scattering above 100 GeV.","keywords":["dark matter","WIMP","neutrino telescope","muon neutrinos","center of the Earth","spin-independent cross section","indirect detection","IceCube"],"falsifier":"Apply the same likelihood to a zenith sideband just outside the signal window (for example 145 degrees to 160 degrees) reweighted by the simulated acceptance; if the best-fit signal fraction or the 90% limits shift by more than the quoted systematic uncertainties, the Monte-Carlo-only background assumption is falsified. A future measurement with more low-energy data that finds a greater than 3 sigma excess at 250 GeV in the bbbar channel would contradict the null result.","tokens_in":1921,"feed_emoji":"🌍","tokens_out":3409,"duration_ms":119712,"temperature":0.7,"pith_summary":"The paper tries to establish whether dark matter particles called WIMPs accumulate and annihilate at the center of the Earth, producing muon neutrinos that a neutrino telescope at the South Pole could see. Using ten years of track-like events (3,266 live days), and testing annihilation into tau+tau-, W+W-, and bbbar at WIMP masses from 10 GeV to 10 TeV, it finds no statistically significant signal. The most signal-like fluctuation is bbbar at 250 GeV with post-trial significance 1.06 sigma, consistent with background. The useful output is a set of 90% confidence upper limits on the spin-independent WIMP-nucleon scattering cross section, which for WIMP masses above 100 GeV are the strongest limits yet set by any neutrino telescope. The result matters because it probes dark matter capture in a different part of the WIMP velocity distribution than direct-detection experiments, so it is a complementary way to rule out dark matter models.","feed_headline":"No dark matter signal from Earth's core in 10 years","feed_subtitle":"If WIMPs annihilate in Earth's center, the telescope sees none; new limits beat all neutrino telescopes above 100 GeV","key_machinery":"The central object is a two-dimensional binned profile-likelihood ratio using the reconstructed zenith angle and reconstructed log-energy of each event as observables, with the signal fraction xi as the parameter of interest and background-component normalizations as nuisance parameters. Signal templates come from simulated WIMP annihilation and propagation; background templates come entirely from Monte Carlo simulations, because the source sits at a fixed local-coordinate position and neither right-ascension scrambling nor off-source regions can provide a data-driven background. The physical link from a measured neutrino flux to a cross section is the capture-annihilation balance dN/dt = C_C - C_A $N^{2}$ - C_E N; for Earth-captured WIMPs equilibrium has not been reached, so limits are quoted in the plane of annihilation cross section <sigma_A v> versus spin-independent scattering cross section $\\sigma$^SI_chiN, and under the canonical thermal annihilation cross section they reduce to limits on $\\sigma$^SI_chiN alone.","core_discovery":"On its own terms, the paper establishes a null result: after 3,266 live days of track-like muon-neutrino events from the direction of Earth's center, testing annihilation into tau+tau-, W+W-, and bbbar at WIMP masses from 10 GeV to 10 TeV, no statistically significant dark matter signal appears. The most signal-like fluctuation is the bbbar channel at 250 GeV, with a post-trial significance of 1.06 $\\sigma$. Interpreting the data as background, the collaboration places 90% confidence upper limits on the spin-independent WIMP-nucleon cross section $\\sigma$^SI_chiN, and for WIMP masses above 100 GeV these are the strongest limits reported by a neutrino telescope to date. Relative to the previous one-year IceCube search, the improvement is greater than a factor of three across the tested masses and channels, reaching about an order of magnitude for m_chi > 100 GeV.","pith_inferences":["Beyond the paper: since the results are quoted as upper limits on event counts, the same likelihood can be recast to constrain any specific model of dark matter interactions; publishing the full likelihood grids would let the community do such recasts without rerunning the detector simulation.","Beyond the paper: the Monte-Carlo-only background is the most fragile link, so a useful cross-check would be to apply the identical selection to the up-going band just outside the signal region (for instance 145 degrees to 160 degrees zenith) as a data-driven background proxy, even though it is not a true off-source sample.","Beyond the paper: combining this Earth-center limit with a Sun-capture limit from the same detector could disentangle spin-independent from spin-dependent WIMP-nucleon coupling, because the Earth's heavy-element composition is dominated by spin-independent scattering.","Beyond the paper: the 1.06 sigma fluctuation at m_chi=250 GeV in the bbbar channel is the concrete feature to watch; if it is a real fluctuation it will fade with more data, while growth would signal the first indirect detection of Earth-captured dark matter."],"forward_implications":["If the null result holds, WIMP-nucleon spin-independent cross sections above roughly 100 GeV are excluded more strongly by neutrino telescopes than before, narrowing the parameter space open to WIMP models.","The limits are the best among neutrino telescopes and are competitive with crystal-based direct-detection experiments, while liquid-xenon experiments remain about an order of magnitude more sensitive.","Because Earth capture favors low-velocity WIMPs and direct detection favors high-velocity recoils, the two search strategies constrain different parts of the WIMP velocity distribution, making the limits complementary rather than redundant.","With more data, all neutrino flavors, or the planned detector upgrade, the same method should reach better sensitivity, particularly for WIMP masses below 100 GeV.","The agreement between data and Monte Carlo at the best-fit points supports the use of simulated backgrounds for this fixed-direction search."],"supporting_citations":[{"why":"the previous one-year IceCube search whose limits this analysis improves by more than a factor of three, and by about an order of magnitude at high mass.","marker":"[25]"},{"why":"supply the simulated neutrino spectra and oscillation-propagated fluxes for each WIMP mass and annihilation channel.","marker":"[37, 38]"},{"why":"provides the previous neutrino-telescope upper limits to which the new limits are compared.","marker":"[22]"},{"why":"generates the atmospheric-muon background that dominates before the event-selection cuts.","marker":"[31]"},{"why":"simulates low-energy neutrino interactions used for the atmospheric background in the low-energy selection.","marker":"[32]"},{"why":"simulates high-energy neutrino interactions used for the atmospheric and astrophysical backgrounds above roughly 100 GeV.","marker":"[33]"},{"why":"supplies the measured astrophysical neutrino flux whose uncertainty band constrains that background component's normalization.","marker":"[36]"},{"why":"defines the profile-likelihood test statistic used to compute significances and 90% confidence upper limits.","marker":"[47]"},{"why":"sets the standard halo model parameters (velocity dispersion, escape velocity, local density) used to convert capture rate into cross-section limits.","marker":"[50]"},{"why":"represents the stronger liquid-xenon direct-detection limit that this analysis is compared against.","marker":"[63]"}],"fun_headline_variants":["10 years of IceCube data: no dark matter from Earth's core","Earth's core yields no WIMP signal in decade of IceCube data","World-leading limits: IceCube's 10-year search for Earth's core dark matter","IceCube finds no dark matter in Earth's core after 10 years"],"cache_read_input_tokens":21632,"weakest_assumption_plain":"The entire search assumes the Monte Carlo simulations of atmospheric muons and neutrinos accurately describe the background in the narrow vertical up-going band around the Earth's center, since the source's fixed position leaves no data-driven way to measure the background from an off-source region.","fun_headline_variants_meta":{"raw":{"variants":["10 years of IceCube data: no dark matter from Earth's core","Earth's core yields no WIMP signal in decade of IceCube data","World-leading limits: IceCube's 10-year search for Earth's core dark matter","IceCube finds no dark matter in Earth's core after 10 years"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000817,"raw_usage":{"total_tokens":3592,"prompt_tokens":971,"completion_tokens":2621,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":587,"completion_tokens_details":{"reasoning_tokens":2539}},"tokens_in":587,"tokens_out":2621,"duration_ms":16809,"temperature":1.0,"reasoning_tokens":2539,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:31:57.201540+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Apply the same likelihood to a zenith sideband just outside the signal window (for example 145 degrees to 160 degrees) reweighted by the simulated acceptance; if the best-fit signal fraction or the 90% limits shift by more than the quoted systematic uncertainties, the Monte-Carlo-only background assumption is falsified. A future measurement with more low-energy data that finds a greater than 3 sigma excess at 250 GeV in the bbbar channel would contradict the null result.","supporting_citations":[{"cited_title":"Albert, et al., Phys","cited_arxiv_id":null,"evidence_quote":"provides the previous neutrino-telescope upper limits to which the new limits are compared."},{"cited_title":"Heck, et al., CORSIKA: A Monte Carlo Code to Sim- ulate Extensive Air Showers(FZKA-6019, 1998)","cited_arxiv_id":null,"evidence_quote":"generates the atmospheric-muon background that dominates before the event-selection cuts."},{"cited_title":"Gazizov, M","cited_arxiv_id":null,"evidence_quote":"simulates high-energy neutrino interactions used for the atmospheric and astrophysical backgrounds above roughly 100 GeV."},{"cited_title":"Cowan, et al., Eur","cited_arxiv_id":null,"evidence_quote":"defines the profile-likelihood test statistic used to compute significances and 90% confidence upper limits."}],"review_version":1}