{"id":"6c251d99-cfd2-4ef4-98e9-600cc52f3c19","arxiv_id":"2510.24114","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Axial coupling of Dirac dark matter to a kinetically-mixed dark photon suppresses the standard direct-detection signal and opens large allowed regions in the WIMP parameter space.","lead":"Dark matter particles that interact with a new 'dark photon' through an axial (spin-dependent) coupling scatter much more weakly off ordinary matter, so their signals are suppressed by the dark matter's low velocity. This makes a whole class of WIMP dark matter models compatible with the relic density, direct detection, and collider bounds over a much wider range of masses and couplings than previously thought.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"LZ limit in Eq. (21) uses inconsistent denominators, so the plotted O8 upper bound may be too weak by up to an order of magnitude; the central allowed regions in Fig. 1 depend directly on this bound.","rationale":"The reader already identified Eq. (21) as internally inconsistent and flagged the simple rescaling of experimental limits as a weakness. My stress test confirms this is the most load-bearing numerical issue: the central allowed-region plots are produced by comparing relic-density lower bounds with directly rescaled LZ bounds, and an order-of-magnitude error in that rescaling can change the conclusion. I do not move the verdict from CONDITIONAL because the O4 cancellation and the O8 operator construction are internally sound, and a corrected recast may still leave a narrower but nontrivial allowed region. I only partially agree with the reader because I would not prioritize the one-mass O8 demonstration as the main blocker; the operator choice is plausible on nuclear-coherence grounds, whereas Eq. (21) is an outright inconsistency in the main quantitative input. The concrete test would settle the issue.","tokens_in":10345,"tokens_out":10200,"duration_ms":95796,"concrete_test":"Recompute the LZ upper bounds on c_p^8 using the correct denominator: take the XENON100 O8 EFT limits from Ref. [43] and rescale by the ratio sigma_SI^p|LZ/sigma_SI^p|XENON100, or better, run a public recast tool such as DDCalc on the LZ 2025 data for the proton-only O8 operator, including the seven xenon isotopes and the isospin-dependent response. Then overlay the thermal-relic c_p^8 lower bounds from Fig. 1 for R=2.05, 2.3, and 3. If the corrected LZ curve lies more than ~0.5 dex below the plotted one, quantify how much of the allowed parameter space survives; if the curve crosses the relic-density band, the central claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative claim that thermal relic density lower bounds lie below LZ direct-detection limits for R=2.05, 2.3, and 3 rests on the LZ O8 bound plotted in Fig. 1. That bound is derived from Eq. (21), which equates (c_p^8)^2|LZ/(c_p^8)^2|Xenon100 to sigma_SI^p|LZ/sigma_SI^p|Xenon1T. The denominator on the left is a XENON100 O8 EFT limit, while the denominator on the right is a XENON1T SI limit; the two are not interchangeable. Since the LZ SI limit is roughly an order of magnitude stronger than XENON1T, and XENON1T is in turn stronger than XENON100, the correct rescaling from XENON100 O8 would use sigma_SI,LZ/sigma_SI,Xenon100 in place of the right-hand ratio. The effect is to lower the LZ curve by about an order of magnitude in (c_p^8)^2, possibly more. Because the right panels of Fig. 1 compare these curves directly with the thermal-relic lower bounds, a downward shift can close or substantially shrink the claimed allowed regions, especially for R=2.3 and R=3. This is a concrete numerical inconsistency in the central figure, not merely a stylistic issue.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a dark sector with a Dirac fermion dark matter particle chi coupled axially to a dark photon A_D that is kinematically mixed with the Standard Model hypercharge boson. After integrating out the heavy states, the low-energy interaction is reduced to non-relativistic effective operators, and the authors argue that O8 dominates direct detection because O4 cancels to leading order and O9 is subleading. For mass ratios R = M_AD/m_chi = 3, 2.3, and 2.05 and dark couplings alpha_D = 0.5, 0.05, 0.005, they set the kinetic mixing epsilon by the thermal relic density and compare the resulting c_p^8 with rescaled XENON100 and LZ limits, and compare epsilon with EWPO and relaxed CMS bounds. They conclude that the allowed parameter space is much broader than in existing vector-portal models.","tokens_in":10802,"tokens_out":7926,"duration_ms":81235,"significance":"If the numerical bounds are reliable, the paper offers an interesting way to evade direct-detection constraints while retaining thermal WIMP dark matter. The O4 cancellation between the dark-photon and Z contributions shown in the Supplemental Material is non-trivial, and the identification of O8 as the leading operator is physically well motivated and potentially useful for future EFT recasts. The paper does not ship code or machine-checked proofs, but the analytic structure of the reduction is clear. The central quantitative claim, however, depends on approximate and, in one place, internally inconsistent rescalings of published direct-detection limits; this is the main obstacle to accepting the stated conclusions.","major_comments":[{"comment":"Equation (21) is internally inconsistent. The left-hand side is the ratio of LZ to XENON100 bounds on (c_p^8)^2, while the right-hand side is the ratio of LZ to XENON1T spin-independent cross-section limits. The denominator on the left is a XENON100 O8 EFT limit and the denominator on the right is a XENON1T SI limit; these are not interchangeable. Since the LZ SI limit is roughly an order of magnitude stronger than XENON1T, and XENON100 is weaker still, the right-hand side is too large, so the LZ curve in Fig. 1 is shifted upward (too weak) by a large factor. The right panels of Fig. 1 are the direct basis for the claim that thermal-relic lower bounds lie below the LZ upper limits for R = 2.3 and R = 3. A corrected rescaling, or preferably a direct recast of LZ data for O8, could close or substantially shrink the claimed allowed regions. This must be fixed before the central result can b","section":"Eq. (21)"},{"comment":"The dominance of O8 over O9 and over their interference is demonstrated only at m_chi = 40 GeV. The relative rate depends on v_min and q^2 through the nuclear response functions, and therefore on m_chi and the target nucleus. The paper scans m_chi from about 1 GeV to 1 TeV, but no analytic scaling argument or additional mass points are provided to justify that O8 remains dominant over the entire scanned range. Since the comparison of the relic-density prediction with the XENON100 and LZ bounds uses only the O8 rate, this assumption is load-bearing and should be checked explicitly.","section":"Supplemental Fig. 3"},{"comment":"The conversion of the XENON100 O8 isoscalar limit to a proton-only limit using a single factor A^2/Z^2 assumes a purely coherent response. For O8 the differential rate involves F_M, F_Delta, F_Sigma', and F_Sigma',Delta, which are not all coherent and have different isospin structure. The authors acknowledge that precise treatment is needed, but because the same approximate limits are used as the exclusion curves in Fig. 1, the statement that the corrections are 'far below the accuracy relevant' should be substantiated, e.g. by showing the size of the isovector contamination for the O8 response. As written, the uncertainty in the direct-detection exclusion curves is not quantified.","section":"Direct detection rescaling"}],"minor_comments":[{"comment":"The text states that the allowed regions are consistent with 'both direct and indirect detection', but no indirect-detection constraints (e.g. Fermi-LAT, HESS, or CTA projections) are computed anywhere in the paper. The abstract later correctly lists only direct detection and collider searches. The Introduction should be tempered or the indirect constraints should be added.","section":"Introduction"},{"comment":"The relaxed CMS bound is shown only for alpha_D = 0.05 and R = 2.3, 2.05. For alpha_D = 0.5 and 0.005 the consistency with collider bounds is asserted but not displayed; a brief statement or additional panel would make the collider claim complete.","section":"Fig. 2"},{"comment":"The phrase 'considerably broader than those found in any existing model' is a strong comparative claim that is not supported by a quantitative comparison with previous models, such as the vector-coupled dark photon or scalar mediator cases. A concrete comparison or a more modest wording would be appropriate.","section":"Summary/Abstract"},{"comment":"The rescaling epsilon < epsilon_CMS * sqrt(Gamma_total/Gamma_SM) is reasonable when the dark photon is produced on shell and decays visibly with reduced branching ratio. The conditions under which this estimate is valid, especially when Gamma_AD->chi chi dominates and the dark photon may be long-lived or invisibly decaying, could be stated more explicitly.","section":"Eq. (22)"}],"recommendation":"major_revision","confidential_remarks":"The main technical issue is Eq. (21), which is not a harmless typo: the right panels of Fig. 1, and therefore the paper's central quantitative conclusion, depend directly on the resulting LZ curve. I would ask the authors to provide either a consistent rescaling or, preferably, a direct EFT recast of the LZ data for O8. The overlap of Ref. [37] with the present authors is not itself a problem, but the EWPO constraints should be described transparently as coming from a related group. The O8 mechanism is interesting enough to warrant a revised version; I would not reject on the basis of the current issues alone."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nTwo things to know up front. First, the axial dark photon portal is a genuinely new twist that keeps Dirac WIMP dark matter alive in a way the vector coupling does not: the leading direct-detection operator becomes O8, and the event rate is suppressed by the DM velocity and momentum transfer. The analytic derivation of the O4 cancellation in the supplement is clean, and the relic-density setup is standard. Second, the central figure (Fig. 1) has a rescaling error for the LZ limit that should be fixed before anyone quotes the allowed regions as exclusions.\n\nThe paper's actual contribution is a parameter-space scan for a Dirac fermion with axial coupling to a kinetically-mixed dark photon. The authors correctly note that the O8 operator is normally neglected in direct-detection analyses, and they show that in this model it dominates. The velocity/momentum suppression is long-established for O8, but applying it to the axial dark photon is new. The result is that for R = 2.05 and 2.3, the thermal-relic lower bound on c_p^8 sits below the XENON100 limit for a wide range of masses. I find that plausible.\n\nThe soft spots are real but not fatal. Equation (21) is internally inconsistent: the left side is LZ/XENON100 for O8 while the right side is LZ/XENON1T for SI. The correct rescaling should compare LZ to XENON100 on both sides. Since LZ is about an order of magnitude stronger than XENON1T, the plotted LZ curve could be too weak by roughly that factor. That does not kill the R=2.05 case, which has plenty of margin, but it could shrink or close the R=2.3 and R=3 regions. The authors should correct this and recompute. Also, the O8 dominance is only demonstrated at m_chi=40 GeV; it is likely fine over the range, but the paper should say it is an assumption. The claim of \"broader than any existing model\" is also not backed by any explicit comparison to other models, so it should be softened.\n\nI'd send this to a serious referee. It is a legitimate phenomenological idea, the calculation is mostly careful, and the flaw is a fixable rescaling issue rather than a wrong mechanism. Once Eq. (21) is corrected and the O8 dominance check is extended to a few masses, it would be a solid Letter.\n\nBest.","headline":"The axial dark photon portal is a real way to rescue Dirac WIMPs, but the LZ rescaling in Fig. 1 is off by up to an order of magnitude and should be corrected before the allowed regions are quoted.","tokens_in":11188,"tokens_out":2954,"would_cite":true,"duration_ms":29868,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.35.+d"],"model":"deepseek-v4-flash","headline":"Dark matter with axial coupling to a dark photon evades direct-detection limits because the dominant scattering operator is suppressed by velocity and momentum transfer.","keywords":["dark matter","dark photon","axial-vector coupling","kinetic mixing","direct detection","thermal relic density","effective field theory","WIMP"],"falsifier":"Direct detection recast: compute the exact O8 event rate for the full xenon isotope composition and nuclear response functions (without rescaling) and compare with public data from current liquid-xenon experiments; if the resulting limit on c_p^8 is more than a factor of a few stronger than the paper's rescaled curve at any mass, the allowed regions shrink significantly. This is a concrete test that can be done with public data.","tokens_in":10204,"feed_emoji":"⚛️","tokens_out":5819,"duration_ms":53920,"temperature":0.7,"pith_summary":"This paper claims that Dirac fermion dark matter with an axial-vector coupling to a kinetically-mixed dark photon produces a non-relativistic direct-detection operator, O8, whose rate is suppressed by the dark matter velocity and momentum transfer. As a result, the standard tension between the thermal relic density and direct-detection upper limits is greatly relaxed: for dark photon masses near the s-channel resonance, the coupling required to match the observed relic abundance falls below current experimental exclusions over dark matter masses from about 1 GeV to 1 TeV. The paper argues that this opens considerably wider regions of the dark parameter space than vector-coupled or standard spin-dependent models allow. A sympathetic reader would care because it revives a technically natural WIMP scenario and gives concrete targets for next-generation direct-detection and collider searches.","feed_headline":"Axial dark photon evades dark-matter limits","feed_subtitle":"Relic-density-compatible WIMPs from 1 GeV to 1 TeV scatter only through a velocity-suppressed operator.","key_machinery":"The central objects are the axial-vector coupling g_chi of the dark photon to Dirac fermion dark matter and the resulting non-relativistic operator O8 (spin of the dark matter dotted into the transverse relative velocity plus half the momentum transfer divided by the reduced mass). The suppression by v_perp and q makes the direct-detection rate parametrically small relative to vector-coupled dark photons. The argument also relies on a nearly exact cancellation between dark-photon and Z-exchange contributions that kills the spin-dependent operator O4, ensuring O8 dominance.","core_discovery":"The central discovery is that the leading effective operator for elastic scattering of this dark matter off nuclei is O8 = S_chi · (v_perp + q/(2 mu)), not the standard spin-independent or spin-dependent operators. Both v_perp and the momentum transfer q are tiny in direct-detection kinematics, so the expected event rate is far smaller than in models with vector coupling. The spin-dependent operator O4 is strongly suppressed by a cancellation between the dark-photon and Z-boson exchange amplitudes, leaving O8 (with a subleading O9) as the dominant channel. The authors show that for mass ratios R = M_AD/m_chi = 2.05 and 2.3, the dark coupling and kinetic-mixing parameter required by thermal f","pith_inferences":["If O8 dominance persists at all masses, a full experimental reanalysis of liquid-xenon data using isospin-asymmetric nuclear response functions would sharpen the bounds; the naive A^2/Z^2 rescaling used here could over- or under-estimate the true limits by a factor related to the neutron-to-proton coupling ratio.","The cancellation that suppresses O4 may be fragile under radiative corrections or UV completion; a one-loop calculation of the spin-dependent coupling would test whether the relic-density-coupling regions survive beyond tree level.","The same velocity/momentum suppression mechanism could be applied to other dark sector models, such as pseudo-Dirac or inelastic dark matter, to relax direct-detection constraints without resonant enhancement.","Future detectors with directional sensitivity or lower thresholds could probe the O8 operator's distinctive q-dependence, providing a distinctive signature that distinguishes this model from spin-independent scenarios."],"forward_implications":["The allowed relic-density-compatible parameter space for WIMP dark matter extends to dark photon masses near the 2 m_chi resonance and dark matter masses up to 1 TeV.","Direct-detection experiments probing O8 will need to account for the v^2 and q^2 suppression; current limits constrain the coupling c_p^8 only weakly.","The kinetic-mixing parameter epsilon in the resonant regime is compatible with electroweak precision observables and with relaxed collider limits when the dark photon decays partly to dark matter.","Near-resonance masses (R approximately 2.05-2.3) provide the broadest allowed regions, whereas R = 3 still conflicts with electroweak precision bounds over a wide mass range.","In this model, existing null results do not exclude a thermal WIMP; the suppression mechanism changes which observables will be most sensitive."],"fun_headline_variants":["Axial dark photon relaxes dark-matter limits","Velocity-suppressed scattering widens WIMP space","O8 operator dodges direct-detection constraints","Axial coupling expands allowed dark-matter mass","Dark photon axial portal eases relic-density bounds"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The analysis assumes that rescaling published direct-detection exclusion limits with a simple A^2/Z^2 factor and a formula that mixes bounds from different experiments accurately reproduces the true O8 constraints across the full mass range.","fun_headline_variants_meta":{"raw":{"variants":["Axial dark photon relaxes dark-matter limits","Velocity-suppressed scattering widens WIMP space","O8 operator dodges direct-detection constraints","Axial coupling expands allowed dark-matter mass","Dark photon axial portal eases relic-density bounds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000176,"raw_usage":{"total_tokens":1071,"prompt_tokens":631,"completion_tokens":440,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":375,"completion_tokens_details":{"reasoning_tokens":368}},"tokens_in":375,"tokens_out":440,"duration_ms":5516,"temperature":1.0,"reasoning_tokens":368,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T07:51:57.141175+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Direct detection recast: compute the exact O8 event rate for the full xenon isotope composition and nuclear response functions (without rescaling) and compare with public data from current liquid-xenon experiments; if the resulting limit on c_p^8 is more than a factor of a few stronger than the paper's rescaled curve at any mass, the allowed regions shrink significantly. This is a concrete test that can be done with public data.","supporting_citations":[],"review_version":1}