{"id":"f168bff3-945a-4113-9347-05ca9a7611ec","arxiv_id":"2412.12286","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"SPHEREx should improve current limits on the axion-photon coupling for 0.5-3 eV ALP dark matter by observing the Milky Way halo in its deep field, the LMC, and dwarf spheroidals.","lead":"This paper forecasts how well the upcoming SPHEREx infrared sky survey could detect the faint light produced when axion-like particle dark matter decays. If the forecast holds, SPHEREx would beat current experimental limits on the axion-photon coupling for axion masses between about 0.5 and 3 eV, mainly by staring at the Milky Way's halo in one of its deep fields.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Foreground subtraction, not the ALP calculation, is the load-bearing step: the MW Deep Field line at the projected limit is ~4e-3 nW/m2/sr against a zodiacal continuum of ~10-100 nW/m2/sr, and a smooth parametric fit may leave solar-spectrum residuals far above photon noise.","rationale":"I checked the order of magnitude of the signal and noise: for the NFW Milky Way halo parameters in Sec. 4.3 and the Deep Field noise quoted in Sec. 3, the 95% C.L. line sensitivity at ma = 1 eV is indeed about 4e-3 nW/m2/sr, corresponding to ga-gamma near 1e-11 GeV^-1, so the particle-physics and halo-modeling part of the calculation is internally consistent. The load-bearing step is not in Eqs. 2.2-4.1 but in the data analysis: the declared plan to fit the continuum with a parametric function and mask bright sources (Sec. 4 after Eq. 4.1). The required dynamic range relative to the zodiacal light is severe, and the uniform morphology of the Milky Way signal gives no spatial handle against a global spectral residual. This is the same assumption the Reader identified, so I agree with the CONDITIONAL verdict. A demonstration on simulated or early SPHEREx data that continuum residuals reach the photon-noise floor would convert the forecast into a firm claim; absent that, the projected blue curve in Fig. 1 should be read as an idealized sensitivity limit rather than a demonstrated reach.","tokens_in":7953,"tokens_out":18150,"duration_ms":181390,"concrete_test":"Use the zodiacal light spectrum at the north ecliptic pole (e.g., from IRTS or the COBE/DIRBE solar-scattering model), degrade it to SPHEREx spectral resolution and channelization, then apply the continuum-fitting procedure described in Sec. 4 (a parametric function fit to the continuum) and record the residual surface brightness in each channel. Compare the maximum residual in the ma=0.5-3 eV channels with the corresponding projected line flux in Fig. 1 (about 4e-3 nW/m2/sr at 1 eV, scaling as m_a^3). If the residual exceeds that line flux, the blue MW curve is not reachable without a high-fidelity zodiacal template and the forecast should be flagged as foreground-limited.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest target, the Milky Way halo in the north Deep Field (Sec. 4.3, Fig. 1), requires detecting a line surface brightness of about 4e-3 nW/m2/sr at ma ~ 1 eV (from Eq. 2.2 and the quoted NFW D-factor, with the 3.8 nW/m2/sr Deep Field noise averaged over ~3.4e6 pixels). The zodiacal light continuum in the same SPHEREx channel is orders of magnitude brighter, of order 10-100 nW/m2/sr, so continuum subtraction must succeed at the 1e-4 relative level. The paper delegates this: 'there are two aspects of the analysis that are not discussed here... background subtraction... masking of bright sources' (Sec. 4, after Eq. 4.1), and the noise figures in Sec. 3 include photon noise from zodiacal light but not residuals from subtracting it. The proposed 'parametric function' continuum fit is not obviously adequate: at R=40-130 the zodiacal spectrum (scattered sunlight) contains solar absorption features, and because the MW-halo ALP line is spatially uniform, a residual spectral feature common to the whole 100 deg2 field cannot be beaten down by pixel averaging. A systematic residual at the 0.1% level, plausible for a low-order parametric fit to a line-blanketed spectrum, would exceed the statistical floor by about 10-100x and remove the claimed improvement over current bounds.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper forecasts the sensitivity of the SPHEREx near-infrared all-sky spectroscopic survey to the two-photon decay line of axion-like-particle (ALP) dark matter. The signal is computed with the standard ALP decay rate and the D-factor formalism of Eq. (2.2), using NFW (and, for the Milky Way, Einasto) halo profiles for eight dwarf spheroidal galaxies, the Large Magellanic Cloud, and the Milky Way halo. The projected 95% C.L. limits on the axion-photon coupling gaγ are derived from a chi-square estimate, Eq. (4.1), using public SPHEREx pixel noise levels for the all-sky and Deep Field surveys. The main result, shown in Fig. 1, is that SPHEREx observations, especially of the Milky Way halo in the north Deep Field, could improve current constraints on gaγ in the 0.5–3 eV ALP mass range.","tokens_in":8262,"tokens_out":8241,"duration_ms":79237,"significance":"If the projected sensitivity is realized, the SPHEREx Deep Field campaign would probe new parameter space for ALP dark matter in the 0.5–3 eV mass range, complementing existing optical and near-infrared line searches (MUSE, WINERED, JWST) and future laboratory experiments such as LAMPOST. The forecast is well posed: no parameter is fitted to the signal, the halo inputs come from published analyses, the instrument noise is taken from public SPHEREx products, and the D-factor calculation is standard. The central quantity (sensitivity to gaγ) is not circularly related to the assumed coupling. The main caveat, explicitly acknowledged in Sec. 4 but not yet quantified, is that the claimed reach depends on the ability to subtract astrophysical continuum foregrounds down to a level far below the photon-noise floor.","major_comments":[{"comment":"The background-subtraction step is load-bearing but is deferred to future work, and I do not think the current text establishes it. For the Milky Way north Deep Field, the projected line surface brightness at the 95% C.L. is of order a few times 10^-3 nW/m^2/sr (from Eq. (4.1) with the quoted Deep Field noise of 2-4 nW/m^2/sr per pixel and roughly 10^6 pixels), while the zodiacal light continuum in the same SPHEREx channel is orders of magnitude brighter. The statement that the continuum will be accounted for by fitting a parametric function is not sufficient, because a residual spectral feature common to the whole ~100 deg^2 field (for example, solar absorption structure in the zodiacal spectrum) cannot be suppressed by spatial pixel averaging. I ask the authors to provide a quantitative assessment: the expected foreground brightness in the relevant channels, the expected residual level after the parametric fit (e.g., from simulations using the SPHEREx spectral resolution), and the corresponding systematic term to be added in quadrature in Eq. (4.1). The same request applies to the LMC and dSph targets, where masking and intrinsic source emission are relevant.","section":"Sec. 4 (Eq. 4.1) and Sec. 4.3"},{"comment":"The noise model is described inconsistently. Section 3 states that the quoted pixel noise includes read noise and the photon noise from zodiacal light, while Sec. 4.2 states that the noise level includes 'the noise induced by the subtraction of the zodiacal light emission.' These are not equivalent: photon noise from a foreground is not the residual after subtracting that foreground. As written, Eq. (4.1) contains only statistical fluctuations and no term for residual systematic contamination. Please state explicitly whether the quoted sigma is purely statistical, and justify that the foreground residual after subtraction is below the statistical floor for each target.","section":"Sec. 3 and Sec. 4.2"}],"minor_comments":[{"comment":"The notation jumps between a photon flux in Eq. (2.2) and a surface brightness in Eq. (4.1). Please specify the explicit conversion factor (including the relevant powers of m_a and hbar) that maps the D-factor to the energy surface brightness in nW/m^2/sr used in the chi-square calculation.","section":"Eqs. (2.1)-(2.2) and Eq. (4.1)"},{"comment":"The text says that ALP decays in the Milky Way halo can contribute significantly to the dSph signal at angular radii near 1 deg, but then states that this contribution is neglected because the signal is spatially uniform. Please clarify whether the Milky Way halo contribution is treated as a background to be subtracted or simply ignored, since a uniform additive component would affect the chi-square if not removed.","section":"Sec. 4.1"},{"comment":"The legend label 'Globular Clusters WINERED' appears to merge two separate constraints (globular clusters and the WINERED survey). Please separate the two entries for clarity.","section":"Fig. 1"},{"comment":"The discussion of an all-sky search assumes the noise scales with the square root of the area and uses 15,000 deg^2 as an example. Please justify this scaling and specify the effective exposure per pixel for the all-sky survey, since the exposure is not uniform over the sky.","section":"Sec. 4.3"},{"comment":"The quoted noise values are for the full two-year nominal mission. For reproducibility, please state the assumed integration time per pixel for the all-sky and Deep Field surveys, or cite the specific public product that provides these noise levels.","section":"Sec. 3"}],"recommendation":"major_revision","confidential_remarks":"This is a straightforward forecasting paper with no controversial physics; the main scientific risk is the foreground subtraction assumption, which is explicitly acknowledged but not quantified. I would support publication after the authors either demonstrate, with a concrete estimate or simulation, that residual foregrounds are subdominant to the photon noise used in Eq. (4.1), or they qualify the abstract and conclusions accordingly. The paper does not supply code, but the calculation is short and reproducible from the text and public products. The paper fits the scope of astro-ph.CO and the journal's interest in forecasts for upcoming facilities."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a clean, useful forecast — the first SPHEREx-specific one for ALP decay lines I know — and the Milky Way north Deep Field target is a genuinely new configuration. But the stress-test concern is right and it is load-bearing: the projected MW-halo line is ~4e-3 nW/m2/sr against a zodiacal continuum of order 10–100 nW/m2/sr, so the continuum subtraction must work at the ~1e-4 relative level. That is not demonstrated, and a smooth parametric fit is unlikely to get there because the zodiacal spectrum contains solar absorption features. Since the MW-halo signal is spatially uniform, those features are common to the whole field and pixel averaging won't remove them. A residual at the 0.1% level would wipe out most of the claimed improvement over current bounds. So the paper's headline claim is conditional in a way the conclusion does not fully convey.\n\nWhat is genuinely good: the calculation is transparent and standard — decay rate, D-factors, SPHEREx noise from public products — and the result doesn't depend on any fitted version of the target coupling. The authors test profile choices and say clearly that masking and background subtraction are deferred. That honesty matters; many forecasts hide those steps. The LMC in the south Deep Field and the MW halo in the north Deep Field are concrete, checkable ideas, and Fig. 1 makes the comparison to existing limits easy to read.\n\nThe soft spots, in proportion: the foreground-subtraction issue is not a minor caveat, it is the main uncertainty. The quoted SPHEREx noise includes photon noise from zodiacal light, but not residuals from modeling and subtracting it. For the LMC, the paper notes the continuum is comparable to zodiacal and says a cleaning procedure should work; that's a hope, not a calculation. Masking bright sources is also deferred, though the expected masked fraction argument is plausible. The dSph forecast is the least interesting but also the least problematic. I would not want to see this paper rejected over the missing foreground study; I would want to see the foreground study added, or at least an explicit mapping of reach versus residual amplitude.\n\nWho is this for: ALP/astro-particle forecasters and anyone planning SPHEREx line searches. It deserves serious peer review. If I were the referee, my main request would be a contamination study for Sec. 4.3, plus a statement of how much of the forecasted gain survives residuals at 0.1% and 1%.","headline":"First SPHEREx ALP forecast with a genuinely clever MW Deep Field target; the headline reach, however, hinges on an unquantified foreground subtraction that may sink it.","tokens_in":8825,"tokens_out":2765,"would_cite":true,"duration_ms":26442,"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 forecasts that SPHEREx, a wide-field near-infrared space telescope, can improve current limits on the axion-photon coupling for axion-like-particle dark matter in the 0.5–3 eV mass range, with the Milky Way halo as the best…","keywords":["axion-like particles","dark matter","SPHEREx","near-infrared spectroscopy","axion-photon coupling","dwarf spheroidal galaxies","Large Magellanic Cloud","Milky Way halo"],"falsifier":"Take real SPHEREx spectra of the north Deep Field, subtract the best-fit continuum model, and measure the residual noise in the spectral channel that would contain the ALP line at $m_a/2$; if the residual exceeds the quoted 2.3 nW/m$^2$/sr per-pixel noise at 5 $\\mu$m, the claimed reach on $g_{a\\gamma}$ is not achievable and must be weakened accordingly.","tokens_in":7728,"feed_emoji":"🔭","tokens_out":11361,"duration_ms":86010,"temperature":0.7,"pith_summary":"This paper forecasts whether the SPHEREx near-infrared all-sky spectral survey can detect the radiative decay of axion-like-particle (ALP) dark matter. If an ALP of mass $m_a$ makes up the dark matter, it decays into two photons, each carrying energy $m_a/2$, producing a narrow spectral line; SPHEREx's 0.75–5 $\\mu$m coverage maps that line to ALP masses of 0.5–3 eV. The authors compute the expected line surface brightness from three targets — dwarf spheroidal galaxies, the Large Magellanic Cloud, and the Milky Way halo — using the integrated dark-matter column (the D-factor) and compare it with SPHEREx's per-pixel sensitivity. They conclude that SPHEREx can significantly improve current bounds on the axion-photon coupling $g_{a\\gamma}$, with the Milky Way halo in the north Deep Field giving the strongest projected sensitivity.","feed_headline":"SPHEREx could beat axion dark-matter limits at 0.5–3 eV","feed_subtitle":"A planned near-infrared all-sky telescope may spot the faint decay line of axion dark matter in the Milky Way halo.","key_machinery":"The central object is the ALP-photon interaction $\\mathcal{L}= -\\tfrac14 g_{a\\gamma} a F_{\\mu\\nu}\\tilde F^{\\mu\\nu}$, which gives ALPs a decay width $\\Gamma_{a\\to\\gamma\\gamma}=g_{a\\gamma}^2 m_a^3/(64\\pi)$. The observable is a nearly monochromatic line at photon energy $E_\\gamma \\simeq m_a/2$, with a flux proportional to the D-factor $D=\\int_{\\Delta\\Omega} d\\Omega \\int_{\\rm l.o.s.} ds\\, \\rho(r)$, the dark-matter column density integrated over the instrument's field of view. The forecast sensitivity is derived from a $\\chi^2$ sum over SPHEREx's 6.2''×6.2'' pixels, comparing the predicted line surface brightness with the quoted per-pixel noise for the all-sky survey and the two 100 deg$^2$ Deep Fields.","core_discovery":"The paper's central claim is that SPHEREx will be able to probe axion-like-particle dark matter in the mass window 0.5–3 eV with sensitivity to the axion-photon coupling $g_{a\\gamma}$ that surpasses existing constraints. The signal is the two-photon decay of ALPs in dark-matter halos, with decay rate $\\Gamma_{a\\to\\gamma\\gamma}=g_{a\\gamma}^2 m_a^3/(64\\pi)$ controlling the flux. For each target, the flux is proportional to a D-factor, the integral of the halo density along the line of sight and over the field of view. Comparing this to the surface-brightness noise in SPHEREx pixels at all-sky and Deep Field depths, the authors find that the Milky Way halo observed in the north Deep Field gives the best reach, because the large sky area compensates for the low dark-matter density away from the Galactic center. The Large Magellanic Cloud, lying partly in the south Deep Field, also improves on current bounds, while the combined dwarf-spheroidal observations give a weaker but complementary constraint.","pith_inferences":["If the foreground subtraction performs as assumed, a line search on real SPHEREx data would also constrain the ALP lifetime fairly directly, since the high-latitude Milky Way D-factor is one of the better-determined astrophysical inputs.","A null result at the projected sensitivity would close part of the favored parameter space for eV-scale ALP dark matter and would strengthen the case for dedicated line-intensity-mapping programs in the same mass range.","The same sensitivity calculation could be applied to other wide-area near-infrared surveys by scaling with sky area and noise, though the paper notes that SPHEREx's all-sky survey does not beat its own Deep Field because of the higher all-sky noise.","A positive line candidate would need to be checked against instrumental artifacts, zodiacal-light residuals, and known emission lines, because the paper does not model those backgrounds in detail."],"forward_implications":["SPHEREx can improve current upper limits on $g_{a\\gamma}$ in the 0.5–3 eV ALP mass range, reaching couplings below the existing optical and near-infrared line constraints.","The Milky Way halo in the north Deep Field is the most sensitive target, with the LMC in the south Deep Field yielding the second-best reach.","A null detection at the projected sensitivity would set new exclusion limits that complement laboratory haloscope searches for masses above about 1 eV.","The same wide-field maps can also be used to search for the extragalactic ALP-decay background through line intensity mapping, with sensitivity comparable to the Milky Way analysis.","Because the bounds scale with the square root of the halo density profile, the forecasts are robust to the choice of NFW versus Einasto profile for the Milky Way and to moderate changes in the LMC density model."],"supporting_citations":[{"why":"Provides the SPHEREx design, wavelength coverage, pixel scale, and per-pixel surface-brightness sensitivities for the all-sky and Deep Field surveys on which the forecasts rest.","marker":"[17–20]"},{"why":"Supplies the Navarro-Frenk-White dark-matter density profile used to model all target halos.","marker":"[25]"},{"why":"Gives the eight dwarf spheroidal galaxies, their tidal radii, and the dark-matter profile parameters that define the dSph forecasts.","marker":"[30]"},{"why":"Supplies the LMC dark-matter density profile and its kinematic extent, which set the ROI and the LMC D-factor.","marker":"[31]"},{"why":"Provides the Milky Way NFW scale radius and normalization used to compute the north Deep Field D-factor.","marker":"[34]"},{"why":"Sets the globular-cluster constraint on the axion-photon coupling that the SPHEREx projections are compared against.","marker":"[16]"},{"why":"Presents an existing near-infrared JWST line search whose exclusion limit the SPHEREx reach aims to surpass.","marker":"[13]"},{"why":"Presents an existing near-infrared WINERED line search that provides another current bound in the same mass range.","marker":"[14]"}],"fun_headline_variants":["SPHEREx to sharpen axion dark-matter limits at 0.5–3 eV","Axion decay line could light up SPHEREx's deep fields","SPHEREx may catch axion photons from dark halos","Near-IR all-sky survey could tighten axion bounds at 0.5–3 eV","SPHEREx targets: axion dark matter in Milky Way halo"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The projected sensitivity assumes that zodiacal light, starlight, and the LMC's own continuum can be modeled and subtracted down to SPHEREx's photon-noise level, and that bright sources can be masked, without leaving a systematic background that mimics or hides the line.","fun_headline_variants_meta":{"raw":{"variants":["SPHEREx to sharpen axion dark-matter limits at 0.5–3 eV","Axion decay line could light up SPHEREx's deep fields","SPHEREx may catch axion photons from dark halos","Near-IR all-sky survey could tighten axion bounds at 0.5–3 eV","SPHEREx targets: axion dark matter in Milky Way halo"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000194,"raw_usage":{"total_tokens":1310,"prompt_tokens":860,"completion_tokens":450,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":476,"completion_tokens_details":{"reasoning_tokens":362}},"tokens_in":476,"tokens_out":450,"duration_ms":4243,"temperature":1.0,"reasoning_tokens":362,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:13:50.379196+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take real SPHEREx spectra of the north Deep Field, subtract the best-fit continuum model, and measure the residual noise in the spectral channel that would contain the ALP line at $m_a/2$; if the residual exceeds the quoted 2.3 nW/m$^2$/sr per-pixel noise at 5 $\\mu$m, the claimed reach on $g_{a\\gamma}$ is not achievable and must be weakened accordingly.","supporting_citations":[],"review_version":1}