{"id":"51c9d024-54f1-4212-be7c-e56bb77f5bc2","arxiv_id":"2505.11607","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"The singlet-doublet dark matter model now passes direct detection limits only with small Yukawa couplings or near blind spots, with coannihilation and a compressed spectrum needed to match the observed relic density.","lead":"This paper re-examines the singlet-doublet dark matter model using the latest LZ direct detection limits. It shows surviving parameter space requires either tiny Yukawa couplings or specially tuned blind-spot regions, and asks whether renormalization group evolution can naturally steer parameters there.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Loop corrections at blind spots may lift direct detection cross sections above the neutrino fog; the imported MSSM-based argument in Sec. II A needs a model-specific check.","rationale":"The reader's verdict is already CONDITIONAL, and agrees on the weakest assumption: tree-level blind spots surviving loop corrections, with the paper importing Ref. [34] rather than computing it for this model. My independent read confirms this as the single most load-bearing concern. The paper's strongest claim is the characterization of the remaining parameter space and the statement that points below the neutrino fog are 'stubbornly out of reach.' That claim would be false if loop corrections restore direct detection cross sections above the fog for the points plotted in Fig. 1. The paper's own text in Sec. II A acknowledges the question and answers it by citing the MSSM-related Ref. [34], but does not exhibit a singlet-doublet calculation. I do not see a reason to change the verdict: the main qualitative conclusions (compressed spectra, coannihilation, small Yukawa or blind-spot regimes) are supported by the scan and analytic structure, but the neutrino-fog-level claim is conditional on a loop calculation that is not shown. The reader's weakest_assumption is identical in substance to my concern, so agreement is 'agree.'","tokens_in":25593,"tokens_out":2379,"duration_ms":22719,"concrete_test":"Compute the one-loop (and, where relevant, two-loop) correction to the SI DM-nucleon cross section in the singlet-doublet model itself, using the model's own spectrum and couplings, for representative benchmarks: (1) the MS < MD h-blind-spot with y1 ~ 0.2; (2) a generic near-blind-spot point with y1 ~ 0.5 and small splitting; (3) the double blind spot with y1 ~ 1. Compare each resulting sigma_SI to the neutrino fog shown in Fig. 1. If sigma_SI exceeds the fog for benchmarks (1) or (2), the 'stubbornly out of reach' claim needs revision; if it stays below, the imported argument from Ref. [34] is validated for this model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central phenomenological claim, that viable parameter space is fully characterized by small Yukawa couplings or blind-spot proximity and that surviving points can sit below the neutrino fog, depends on the assertion in Sec. II A that loop corrections do not spoil the blind-spot suppression. The paper does not compute these loop corrections for the singlet-doublet model; it imports Ref. [34], which demonstrated blind-spot persistence in the MSSM context. This is an external transfer of a quantitative result, and the singlet-doublet model differs from the MSSM in field content and couplings (no gauge-kinetic normalization tying y1, y2 to g', additional light states, and a different Higgsino/Bino mixing structure). In particular, the Z boson blind spot at y2/y1 = -1 is protected by a custodial symmetry, so loop-induced SI scattering there may genuinely remain below the neutrino fog; however, the h blind spot for MS < MD and the generic near-blind-spot points in Fig. 1 involve accidental cancellations not protected by a symmetry, and loop-level corrections (two-loop Higgs exchange, box diagrams, and inelastic transitions) could restore a cross section above the neutrino fog. The claim that points beneath the fog are 'stubbornly out of reach' is the strongest conclusion, so this assumption is load-bearing. If the loop-level cross section is larger than the fog for some of these points, the paper's headline characterization of the remaining parameter space would shrink qualitatively.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript revisits the singlet-doublet fermion dark matter model in light of the 2024 LZ direct detection results. It scans the four-dimensional parameter space (MS, MD, y1, y2) with SARAH, SPheno and micrOMEGAs, imposing the Planck relic density, thermalization of the dark sector, and LZ bounds. The authors find that the surviving parameter space is organized into two regimes: small Yukawa couplings, or proximity to the h, Z, or double blind spots, with coannihilation and a compressed mass spectrum required to obtain the observed relic abundance. The paper then characterizes the mass spectrum, the dominant early-universe processes, LHC production and decay signatures, and the possibility of RG evolution focusing generic high-scale parameters into the blind-spot regions. Appendix A provides analytic diagonalization formulas and explicit couplings in the small-Yukawa and blind-spot limits, and the scan data are made publicly available.","tokens_in":25875,"tokens_out":7887,"duration_ms":86903,"significance":"If the main claims hold, this is a valuable and timely phenomenological update: it sharply characterizes the remaining viable parameter space of a well-motivated WIMP model, identifies a region below the neutrino fog that will be difficult to probe, and gives concrete LHC signatures that follow from the compressed spectrum. The paper is transparent in its use of standard public codes, provides analytic expressions for the relevant couplings, and releases the scan data, all of which strengthen its usefulness. The RG-focusing discussion is an interesting addition, although the authors appropriately note that the most attractive focusing scenario is in tension with gauge coupling unification. The most consequential claim, that some surviving points remain 'stubbornly out of reach' of direct detection, is also the one that rests most heavily on an imported loop-level argument.","major_comments":[{"comment":"The assertion that loop corrections do not spoil the blind-spot suppression, and that loop-induced direct detection rates remain below the neutrino fog, is imported from Ref. [34] in the MSSM context rather than demonstrated for the singlet-doublet model. This is load-bearing for the paper's headline conclusion that points beneath the neutrino fog are 'stubbornly out of reach.' The Z blind spot at y2/y1 = -1 is protected by a custodial symmetry, but the h blind spot of Eq. (12) (MS < MD) and the generic near-blind-spot points are accidental cancellations, and the singlet-doublet model differs from the MSSM in field content and loop structure. I request a model-specific estimate, ideally an explicit one-loop computation of the spin-independent cross section for representative surviving points, or a clearly softened claim that the below-fog region is a tree-level statement whose loop-level viability remains to be checked.","section":"Sec. II A, after Eq. (13), with Figs. 1 and 3 and the Conclusion"},{"comment":"The statement that all viable parameter space requires coannihilation, and hence a compressed spectrum, is central to the paper's two-regime summary, but the main scan restricts the mass ratio to 0.7 < MS/MD < 1.2. This range by construction preferentially selects compressed spectra. The text mentions additional exploratory scans that found no points without coannihilation, but gives no details on their coverage or robustness. Since the paper claims that the remaining parameter space is 'fully characterized,' please either provide details of the exploratory scans or rephrase the claim to be explicitly about the surveyed region.","section":"Sec. III A, scan range and text near Fig. 2 and the Conclusion"},{"comment":"The relic density calculation uses micrOMEGAs with the fast option and with off-shell gauge-boson final states neglected (VW/VZdecay=0). For the low-mass part of the scan and for compressed spectra near thresholds, this approximation can shift the computed relic density by an amount comparable to the 10% tolerance used to select surviving points. Please quantify the impact of this approximation on the displayed relic-density curves and on the boundaries of the surviving regions, or justify that all retained points are sufficiently far from the affected thresholds.","section":"Sec. III A, footnote 1 and relic density discussion"}],"minor_comments":[{"comment":"There are several typographical errors, including 'coannhilation' in Secs. II A and III C and 'direction detection' in Sec. II A; these should be corrected.","section":"Throughout"},{"comment":"The unsymmetrized nature of the Higgs coupling gh is easy to miss; consider adding a sentence in the text immediately after Eq. (10) clarifying how the couplings are symmetrized when inserted into the interaction Lagrangian for i ≠ j.","section":"Eqs. (9)-(10)"},{"comment":"The phrase 'beyond the LZ projected sensitivity' is ambiguous: in the context of cross sections it means 'below the projected sensitivity curve,' and I suggest rewording the captions to avoid confusion.","section":"Sec. III B, captions of Figs. 6 and 7"},{"comment":"The tension between the large-Yukawa focusing scenario and gauge coupling unification is an important caveat; I suggest moving it from the footnote into the main text of Sec. IV 2.","section":"Sec. IV 2, footnote 3"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid phenomenological update with clean analytic apparatus and useful public data. The main reservation is that the below-fog conclusion depends on an unverified transfer of a loop-level result from the MSSM to the singlet-doublet model; this is fixable either by a dedicated computation or by a softened claim. The coannihilation-universality claim also needs qualification given the deliberately narrow scan range. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a serious, well-executed update of a classic WIMP model, and the central two-regime picture is probably right; but the paper leans on an imported MSSM result for loop corrections to blind spots, and the RG discussion is less \"generic\" than the abstract claims.\n\nThe genuinely new pieces are the LZ-2024 update of the surviving parameter space, the compressed-spectrum mass predictions (especially Fig. 7), and the RG analysis of why the mass ratio might focus toward the double blind spot. The scan is large and uses standard tools (SARAH/SPheno/micrOMEGAs), the analytic diagonalization in Appendix A is clean, and the authors make the surviving-point dataset publicly available. Those are real strengths, and the citations to prior blind-spot work are fair. The claim that all remaining viable points require coannihilation is well supported by their scan, with the caveat that the scan range was deliberately centered on MS/MD near one.\n\nThe main soft spot is the loop-correction discussion in Sec. II A. The paper asserts, with Ref. [34] from the MSSM, that loop corrections do not spoil blind-spot suppression and that the loop-induced SI cross section sits below the neutrino fog. But the h blind spot for MS<MD is an accidental cancellation, not a symmetry-protected one. The singlet-doublet model differs from the MSSM in couplings and field content, so the MSSM calculation is not a substitute. If loop corrections restore a cross section above the fog, the \"stubbornly out of reach\" points in Fig. 1 shrink. This does not kill the two-regime classification or the coannihilation/compressed-spectrum message, but it does weaken the sharpest conclusion, so it deserves a model-specific one-loop calculation.\n\nThe RG section is honest about the stability of ry=±1, but the abstract's \"generic ultraviolet initial conditions\" oversells it. Focusing to the double blind spot requires ry=-1 at the high scale; small perturbations are IR-repulsive, as the paper itself shows. That should be softened to \"a class of UV initial conditions.\" The coannihilation-necessity claim is also partly shaped by the scan range; the paper says exploratory scans found nothing, which is reassuring, but a targeted analytic argument would be better.\n\nBottom line: worth a serious referee. The paper is a useful reference for DM phenomenologists and for LHC searches targeting compressed electroweak states. I'd send it to review with a request to add a loop-level check of the blind-spot region and to fix the RG wording.","headline":"Solid LZ-2024 update of singlet-doublet DM with useful compressed-spectrum predictions; main caveats are an imported MSSM loop-correction argument and an RG claim that is less generic than advertised.","tokens_in":26419,"tokens_out":3908,"would_cite":true,"duration_ms":41955,"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":"After the 2024 LZ limits, the singlet-doublet dark matter model survives only with tiny Yukawa couplings or near a blind spot, and in both cases requires coannihilation and a compressed spectrum.","keywords":["singlet-doublet dark matter","WIMP","direct detection blind spots","coannihilation","compressed mass spectrum","renormalization group focusing","LHC dark matter searches","neutrino fog"],"falsifier":"Compute the one-loop correction to the spin-independent cross section at the Higgs and double blind spots for representative viable points (for example $M_S=300$ GeV with $y_1=0.25$, or $M_S=M_D=850$ GeV). If the loop result exceeds the projected direct detection sensitivity at those masses, the region claimed to lie below the neutrino fog disappears. A second check is experimental: search 13.6 TeV collider data for the predicted soft-lepton or pion final states from $χ^\\pm\\toχ^0_1$ and $χ^0_2\\toχ^0_1$ decays; an exclusion stronger than the model's cross-section curves would rule out the surviving compressed-spectrum points.","tokens_in":25346,"feed_emoji":"⚛️","tokens_out":8573,"duration_ms":79157,"temperature":0.7,"pith_summary":"The paper asks what survives of the singlet-doublet WIMP model after the most recent direct detection bounds. It claims that the only ways to evade those bounds are to make the Yukawa couplings very small or to sit on a 'blind spot' where the dark matter's coupling to the Higgs boson, the $Z$ boson, or both vanishes. In either regime the dark matter annihilates too feebly by itself, so the observed relic abundance is set by coannihilation with heavier dark-sector states, which forces a compressed mass spectrum. The surviving points include direct detection cross sections below the projected sensitivity of upcoming experiments and even below the neutrino fog, where dark matter signals are indistinguishable from neutrino backgrounds. Collider signatures are then sharply predicted: production of near-degenerate doublet states decaying through off-shell $W$ and $Z$ bosons.","feed_headline":"Singlet-doublet dark matter is squeezed into two hiding places","feed_subtitle":"Both regimes force coannihilation and compressed mass spectra, giving collider searches a clear target.","key_machinery":"The engine is the $3\\times3$ neutral mass matrix $M_N$ built from $M_S$, $M_D$, $y_1$, and $y_2$, whose diagonalization gives three Majorana fermions and one charged Dirac fermion. Direct detection at tree level is controlled by two couplings: the DM--Higgs coupling $g_{h\\chi^0_1\\chi^0_1}$, which generates the spin-independent signal, and the DM--$Z$ coupling $g_{Z\\chi^0_1\\chi^0_1}$, which generates the spin-dependent signal. The blind spots are the conditions under which these couplings vanish: $y_2/y_1=\\pm1$ for the $Z$ blind spot, the Higgs blind spot relation of Eq. (12) for $M_S<M_D$, and the double blind spot $y_2/y_1=-1$ for $M_D<M_S$. The second pillar is coannihilation, whose Boltzmann suppression $\\sim e^{-x\\Delta_i}$ with $\\Delta_i=(m_i-m_1)/m_1$ makes the relic density sensitive to the fractional mass splittings, so a compressed spectrum is a consequence of the constraints rather than an assumption. The third is a one-loop renormalization group analysis of the ratios $r_m\\equiv M_S/M_D$ and $r_y\\equiv y_2/y_1$, which shows that $r_y=-1$ is a fixed point of enhanced symmetry and that $r_m$ can flow toward a pseudostationary point $r_m^*$, focusing parameters into the double blind spot in the infrared.","core_discovery":"The paper establishes that consistency with the observed relic density and the 2024 direct detection bounds forces the singlet-doublet model into one of two corners. In the first, the Yukawa couplings are so small that the dark matter's direct detection cross section is automatically tiny. In the second, the parameters sit near a tree-level blind spot where the DM--Higgs coupling $g_{h\\chi^0_1\\chi^0_1}$ vanishes, the DM--$Z$ coupling $g_{Z\\chi^0_1\\chi^0_1}$ vanishes, or both vanish at the 'double blind spot' $y_2/y_1=-1$ with $M_D<M_S$. In both corners the dark matter's own annihilations are too weak to set the abundance, so the observed relic density is achieved by coannihilation with neighboring dark-sector states; this forces the spectrum to be compressed, with one state near $M_S$ and three near $M_D$. For $M_S\\lesssim 850$ GeV the dark matter is singletlike with mass splittings of order 1--20 GeV, while for larger masses it is doubletlike and nearly degenerate with the charged state. The paper finds points whose direct detection cross sections fall below the projected sensitivity of LZ and even below the neutrino fog, and it shows that those hard-to-detect regions have specific LHC signatures. It also argues that renormalization group evolution can focus a range of ultraviolet mass ratios onto the double blind spot when the Yukawa couplings are large, making that corner less accidental.","pith_inferences":["Editorial inference: the claim that blind-spot points remain below the neutrino fog leans on a loop-correction result imported from the MSSM; a dedicated one-loop calculation of the singlet-doublet spin-independent cross section at the blind spot is the direct test of this corner, and if loops restore a larger cross section, the 'stubbornly out of reach' region shrinks.","Editorial inference: because both surviving regimes produce near-degenerate doubletlike states with electroweak pair-production cross sections, collider searches for soft leptons and short track gaps can in principle cover the blind-spot region that direct detection cannot, making colliders the decisive experiment for this model.","Editorial inference: the same blind-spot-plus-coannihilation logic transfers to neighboring models such as singlet-triplet and Wino-Bino-like dark matter, and the RG-focusing analysis indicates which of those corners can emerge from generic ultraviolet conditions and which require coincidence."],"forward_implications":["Every remaining viable point of the singlet-doublet model has a compressed dark sector, with the charged state and second neutral state within roughly 0--20 GeV of the dark matter, so future WIMP searches should target compressed spectra rather than isolated heavy states.","Direct detection cannot close the remaining parameter space: points below the neutrino fog evade even the projected sensitivity of upcoming experiments, so ruling out the model requires complementary probes.","At the LHC, dark-sector pair production cross sections match the pure doublet limit, and the produced states decay through off-shell $W$ and $Z$ bosons to soft leptons; some small-Yukawa corners give displaced vertices.","Large Yukawa couplings near the double blind spot allow dark matter masses up to about 1500 GeV, beyond the roughly 1100 GeV limit set by gauge interactions alone, which extends the mass range that collider searches must cover.","Renormalization group focusing can make the double blind spot natural: with $r_y=-1$ in the infrared and $y_1\\gtrsim 0.6$, ultraviolet values with $M_S<M_D$ evolve to $M_S\\gtrsim M_D$ at low energies, placing the model directly in the blind-spot regime."],"supporting_citations":[{"why":"Supplies the 2024 LZ exclusion limits that define the surviving parameter space.","marker":"[1]"},{"why":"Derives the Higgs and Z blind-spot conditions (Eqs. 11--13) around which the viable scan points cluster.","marker":"[4]"},{"why":"Provides the coannihilation formalism and the Boltzmann-suppression factor used to connect relic density to mass splittings.","marker":"[30]"},{"why":"Establishes in the MSSM that loop corrections shift rather than destroy blind spots, underpinning the claim that loop-induced cross sections stay below the neutrino fog.","marker":"[34]"},{"why":"Gives the projected LZ sensitivity against which surviving points are classified as beyond future direct detection reach.","marker":"[46]"},{"why":"Defines the neutrino fog used to identify points that remain stubbornly hard to probe.","marker":"[47]"},{"why":"Introduces the renormalization group focusing framework applied to the singlet-doublet mass and coupling ratios.","marker":"[63]"}],"fun_headline_variants":["Singlet-doublet dark matter: two hiding places left","Two corners remain for singlet-doublet dark matter","Coannihilation corners: where dark matter still fits","Blind spot and tiny coupling: dark matter's last refuges","RG focus lands dark matter on the double blind spot"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The tree-level blind-spot cancellation survives loop corrections, so loop-induced direct detection rates stay below the neutrino fog; the paper imports this result from the related MSSM context rather than computing the loop correction in the singlet-doublet model itself.","fun_headline_variants_meta":{"raw":{"variants":["Singlet-doublet dark matter: two hiding places left","Two corners remain for singlet-doublet dark matter","Coannihilation corners: where dark matter still fits","Blind spot and tiny coupling: dark matter's last refuges","RG focus lands dark matter on the double blind spot"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000692,"raw_usage":{"total_tokens":3123,"prompt_tokens":927,"completion_tokens":2196,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":2113}},"tokens_in":543,"tokens_out":2196,"duration_ms":17629,"temperature":1.0,"reasoning_tokens":2113,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:52:54.466633+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the one-loop correction to the spin-independent cross section at the Higgs and double blind spots for representative viable points (for example $M_S=300$ GeV with $y_1=0.25$, or $M_S=M_D=850$ GeV). If the loop result exceeds the projected direct detection sensitivity at those masses, the region claimed to lie below the neutrino fog disappears. A second check is experimental: search 13.6 TeV collider data for the predicted soft-lepton or pion final states from $χ^\\pm\\toχ^0_1$ and $χ^0_2\\toχ^0_1$ decays; an exclusion stronger than the model's cross-section curves would rule out the surviving compressed-spectrum points.","supporting_citations":[{"cited_title":"Dark Matter in the Finely Tuned Minimal Supersymmetric Standard Model","cited_arxiv_id":"hep-ph/0406144","evidence_quote":"Gives the projected LZ sensitivity against which surviving points are classified as beyond future direct detection reach."}],"review_version":1}