{"id":"3885db21-1081-4e3a-97b4-e1e91bd6f699","arxiv_id":"1908.11387","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Projected MATHUSLA, HL-LHC, and FCC-hh forward detector sensitivities probe Higgs-mediated freeze-in dark matter across parent masses up to about 10 TeV.","lead":"This paper maps how proposed long-lived particle detectors could observe dark matter produced by the freeze-in mechanism, using a Higgs-mediated singlet-doublet model as a test case. It finds that MATHUSLA and a future 100 TeV forward detector could cover much of the cosmologically allowed parameter space that prompt LHC searches miss.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"MATHUSLA/FCC-hh reach rests on idealized detector assumptions, notably an unvalidated acceptance near m2 ~ mh + m1 and a background-free forward detector; a detector-level simulation would test the claimed coverage.","rationale":"The reader's weakest_assumption correctly identifies the idealized detector treatment as the main soft spot. I agree: the projections in Sections 5 and 7 set 95% CL limits with three background-free signal events and assume perfect efficiency, while the MATHUSLA acceptance near m2 ≈ mh + m1 is explicitly left unvalidated in Section 5. This is the single most load-bearing concern because the headline claims about probing DM masses up to 1–10 GeV with MATHUSLA and parent masses up to 10 TeV with a forward FCC-hh detector follow directly from these assumptions. The paper itself acknowledges the mono-jet extrapolation is overly aggressive and the DV recast is less trustworthy in the compressed region, so those are already flagged. No internal inconsistency in the thermal-mass treatment or the freeze-in calculation was found; the factor-of-five effect is physically plausible and is bracketed by the TEW variation. The proposed check—simulating the decay-product hit rate inside MATHUSLA for the non-collinear regime—would settle the acceptance concern quantitatively. Since the paper's conclusions are already framed as projections and the reader's conditional verdict accounts for these uncertainties, the verdict should remain CONDITIONAL (i.e., unchanged by this stress-test).","tokens_in":29573,"tokens_out":18875,"duration_ms":186587,"concrete_test":"Generate χ2 → h χ1 events for benchmark m2 = 130 GeV, m1 = 1 GeV with MadGraph/Pythia, force the χ2 to decay uniformly along the MATHUSLA volume with positions sampled from the decay law used in Eq. (5.3). For each event where the χ2 trajectory intersects the tracking layers, record whether the final-state particles from h (after hadronization) yield at least two tracks with pT > 1 GeV that also intersect the tracking layers. Compute the average 'collinearity acceptance' over the volume. If it is below ~0.8, rescale the signal efficiency used in Figures 6 and 9 and check whether the DM mass reach remains above m1 = 1 GeV; a significant drop would confirm that the unvalidated acceptance assumption is load-bearing for the headline claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central sensitivity claims in Sections 5 and 7 depend on two optimistic detector assumptions. First, Section 5 states: 'we assume for simplicity that the visible χ2 decay products would then also hit the tracking layers if the decay happens inside MATHUSLA (for m2 ≫ mh + m1 this occurs automatically... while for m2 → mh + m1 addressing the possible modification of our acceptances requires a more detailed event and detector simulation beyond the scope of this work).' This unvalidated collinearity assumption matters precisely in the region m2 ≈ 126–135 GeV that contributes to the claimed reach in DM mass up to m1 ~ 1–10 GeV (Figure 6). If the true acceptance is lower, the sensitivity contours shrink and the 'wide region' claim weakens. Second, Section 7 assumes 'perfect detector performance as well as a background-free environment' for the FCC-hh forward detector volume (z ∈ [20, 40] m, ρ ∈ [5, 30] m) and sets the 95% CL reach at Nevents = 3. No background estimate is provided for neutral hadron interactions, pile-up, or fake displaced vertices in this forward region at a 100 TeV pp collider; if backgrounds are not negligible, the 10 TeV parent-mass reach and the BBN-limit coverage for m2 ≲ 600 GeV are not established. These assumptions are load-bearing because they directly convert production cross sections into exclusion limits without a demonstrated handle on backgrounds or reconstruction efficiency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies freeze-in dark matter production through the decay of a neutral parent particle, using as a concrete model a singlet-doublet fermion extension of the Standard Model with the Higgs as the visible decay product. It computes the dark matter relic density with modified micrOMEGAs code including thermal masses and a temperature-dependent treatment of electroweak symmetry breaking, and it derives cosmological constraints from Big Bang Nucleosynthesis and Lyman-alpha observations. The main phenomenological results are projected 95% C.L. sensitivities for the MATHUSLA100/MATHUSLA200 surface detectors, for ATLAS/CMS searches at the HL-LHC (mono-jet, disappearing tracks, and displaced vertices plus missing transverse energy), and for a forward detector at a future 100 TeV FCC-hh. The authors find that MATHUSLA can probe a wide region between the Lyman-alpha and BBN constraints, that the FCC-hh forward detector could extend the reach to parent masses up to about 10 TeV, and that thermal-mass corrections can change the freeze-in relic abundance by up to a factor of about five in parts of the parameter space.","tokens_in":29875,"tokens_out":7339,"duration_ms":73614,"significance":"If the projections hold, the paper provides a useful and fairly complete case study of how long-lived particle searches can probe freeze-in dark matter, including a well-documented recast of the ATLAS displaced-vertex search and the use of NLO+NLL Higgsino production cross sections. The treatment of thermal masses and of the electroweak phase transition addresses a gap in earlier freeze-in studies, and the comparison among MATHUSLA, LHC searches, and the FCC-hh forward detector gives a clear picture of complementarity. The authors are also unusually explicit about the limitations of their assumptions, which makes the paper easier to evaluate. The headline sensitivity claims, however, rest on idealized detector assumptions and on a benchmark electroweak transition temperature that is not realized in the minimal model, so the quantitative conclusions should be treated as conditional.","major_comments":[{"comment":"The MATHUSLA projections are obtained by assuming that visible decay products of chi2 hit the tracking layers whenever the chi2 trajectory does, and the text explicitly states that this is only automatic for m2 much larger than mh + m1. For m2 approaching mh + m1 the authors acknowledge that a more detailed event and detector simulation is needed. This is a load-bearing assumption because Figure 6 and the conclusions use it to claim coverage of a wide region of the freeze-in parameter space, including parts where m2 is not large compared with mh. The authors should either provide a detector-level or conservative geometric treatment of this region or explicitly restrict the coverage claim to the regime where collinearity holds.","section":"Section 5, Eqs. (5.2)-(5.4), Fig. 6"},{"comment":"The FCC-hh forward detector sensitivity is computed under the assumptions of perfect detector performance and a background-free environment, with the 95% C.L. reach set at three signal events, but no background estimate is given for the forward volume z in [20,40] m, rho in [5,30] m at a 100 TeV pp collider. Neutral hadron interactions, pile-up, and fake displaced vertices could produce backgrounds in this geometry, and the central claims of reaching parent masses up to 10 TeV and lifetimes up to the BBN bound for m2 below about 600 GeV are derived directly from this assumption. A background estimate or a well-justified background-free argument is needed before these projections can be considered robust.","section":"Section 7, Eq. (5.4) and Fig. 10"},{"comment":"The recast validation shows that the authors' derived limits are stronger than those of ATLAS in the compressed gluino region, and the text acknowledges that the corresponding freeze-in region is m2 close to mh. The DV + Emiss_T exclusion contours in Figure 9 therefore overestimate the true sensitivity precisely where the Higgs decay channel is phase-space suppressed. Since these contours are used in the comparison with MATHUSLA and in the complementarity discussion, this region should be marked as approximate or the recast should be improved.","section":"Appendix B, Figs. 13-14 and Fig. 9"},{"comment":"The statement that thermal masses change the relic density by up to a factor of about five relies on the benchmark TEW = 50 GeV, which the authors describe as corresponding to a strongly super-cooled first-order electroweak phase transition. The model defined by Eq. (3.2) contains no additional scalar dynamics that would turn the Standard Model crossover into such a transition, so the largest thermal-mass effect is not realized in the minimal model. The authors should either justify how TEW = 50 GeV arises within the model or present the factor-five result as an illustration for a non-minimal extension rather than as a property of the specific freeze-in scenario studied here.","section":"Section 3.2, Figs. 2-3"}],"minor_comments":[{"comment":"The decay-length estimate uses '4 Km'; for consistency with the rest of the paper and the figures, this should be '4 km'.","section":"Section 5, Eq. (5.1)"},{"comment":"The mono-jet rescaling uses up-type quark couplings only, and the text notes that a full treatment would re-weight down-type contributions; this approximation should be stated directly in the equation or its caption so that readers do not mistake Eq. (6.1) for the exact signal strength.","section":"Section 6.1, Eq. (6.1)"},{"comment":"The caption lists branching-ratio values but does not explain how to read the color scale or contours in the (m2, ctau_chi2) plane; a brief description would improve interpretability.","section":"Figure 1, right panel"},{"comment":"The geometric acceptance epsilon_geometric ~ 0.5 for the forward detector is quoted without a derivation or a definition of whether it includes the decay-product geometry; stating the definition used in the simulation would make the result reproducible.","section":"Section 7"}],"recommendation":"major_revision","confidential_remarks":"This is a solid phenomenological projection paper with a clear framework and unusually honest caveats. The reason for major revision is that the headline sensitivity claims and the factor-five thermal-mass statement rest on assumptions that are either explicitly unvalidated (MATHUSLA acceptance near m2 ~ mh + m1), unjustified for the forward detector (background-free environment at 100 TeV), or not realized in the minimal model (TEW = 50 GeV). These are fixable within the scope of the paper by adding conservative estimates, background arguments, or explicit qualifications, but they are load-bearing rather than cosmetic. The paper is well within the scope of JHEP."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a careful phenomenological paper that does one genuinely new thing—it computes freeze-in relic densities from Higgs decay including thermal masses and a finite EW phase-transition temperature—and then maps the resulting viable parameter space onto MATHUSLA, HL-LHC, and an FCC-hh forward detector. The sensitivity study is thorough, the ATLAS displaced-vertex recast is validated against the published gluino limits, and the authors are honest about where their approximations break down. The central conclusions are plausible, but the exact reach numbers should be quoted with the detector idealizations attached.\n\nWhat is actually new: prior singlet-doublet freeze-in studies (Calibbi et al.) did not include thermal masses or the EW phase transition in the relic calculation. Here that changes the relic density by up to a factor of about five in parts of parameter space, and it matters for the correct curve in the (m2, cτ) plane. The dedicated MATHUSLA100/200 sensitivity maps and the 100 TeV forward-detector projection are also new for this model. The paper is well organized and the technical appendices are useful; the NLO+NLL cross sections from Resummino and the CLs treatment of the 3000 fb−1 DV projection are appropriate.\n\nSoft spots, in proportion: the sensitivity curves in Sections 5 and 7 assume background-free detectors, perfect efficiency, and three signal events for 95% CL. The authors say this explicitly, but it means the reach is an idealized ceiling, not a realistic projection. The more specific concern is the MATHUSLA acceptance assumption: for m2 near mh + m1 the decay products are not necessarily collinear with the χ2 trajectory, and the authors acknowledge that validating this needs a detector simulation. That region contributes to the claimed DM mass reach up to about 1–10 GeV, so the 'wide region' wording is a bit stronger than the analysis strictly supports. The mono-jet projection to 300 fb−1 is flagged by the authors themselves as overly aggressive, so I wouldn't put any weight on that curve. The DV recast is also less trustworthy near m2 ~ mh + m1; again acknowledged in the appendix. No code or data files are released, so an independent group would have to reimplement to reproduce the curves—common for 2019 phenomenology but worth noting.\n\nOverall: the load-bearing assumptions are stated, not hidden, and the qualitative conclusions—MATHUSLA can probe lifetimes beyond the LHC's reach, a forward detector extends the mass reach dramatically—are robust to the idealizations. The factor-five thermal correction is the kind of thing that should be checked by other groups, but the logic is sound.\n\nWho it's for: anyone working on freeze-in models or LLP detectors. I'd send it to peer review; the authors have done the community a service by putting the thermal-mass issue on the table. I'd cite it if I were working on LLP phenomenology.","headline":"Solid, transparent freeze-in LLP study with real technical progress in the relic calculation; read the sensitivity curves with the stated detector idealizations in mind.","tokens_in":30393,"tokens_out":2075,"would_cite":true,"duration_ms":19651,"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":"The paper argues that a surface detector at the LHC could cover a wide region of viable freeze-in dark-matter parameter space, probing dark-matter masses up to about 1–10 GeV, and that a forward detector at a 100 TeV collider could reach…","keywords":["freeze-in dark matter","long-lived particles","MATHUSLA","Higgs portal","displaced vertices","electroweak phase transition","thermal masses","Lyman-alpha bound"],"falsifier":"A full detector-level simulation of MATHUSLA including cosmic-ray and beam backgrounds, applied to the χ2 → h χ1 signal, would settle the reach: if backgrounds exceed about one event in the signal region at 300 fb⁻¹, the three-signal-event sensitivity contours in Figures 6 and 9 do not hold, and simulating decays with m2 close to mh + m1 would test whether the acceptance assumption that decay products hit the tracking layers holds.","tokens_in":29336,"feed_emoji":"🔭","tokens_out":7455,"duration_ms":66017,"temperature":0.7,"pith_summary":"Freeze-in dark matter is produced so feebly that the parent particles decaying into it can be long-lived, and this paper asks whether proposed long-lived-particle detectors could catch those decays. Using a minimal Higgs-portal model in which a fermion singlet mixes with a fermion doublet after electroweak symmetry breaking, it computes the relic density with thermal masses and the electroweak phase transition included, and then projects the reach of MATHUSLA, HL-LHC displaced-vertex searches, and a forward detector at a 100 TeV collider. The central claim is that MATHUSLA could cover a wide slice of the cosmologically allowed freeze-in parameter space between the Lyman-alpha and BBN limits, probing dark-matter masses up to about 1–10 GeV, while the 100 TeV forward detector could probe parent masses up to about 10 TeV. A sympathetic reader would care because freeze-in is one of the few non-thermal dark-matter production mechanisms that leaves a concrete, testable collider imprint, and the projected sensitivities turn a cosmological production mechanism into an experimental target.","feed_headline":"A surface detector could catch freeze-in dark matter up to 10 GeV","feed_subtitle":"Long-lived Higgs-portal decays would connect colliders to the Lyman-alpha-to-BBN gap.","key_machinery":"The load-bearing object is the long-lived parent state χ2, a mostly-doublet Dirac fermion whose mixing with the dark-matter singlet is set by yχv/√2(m2 − m1). The production mechanism is freeze-in from decay: bath particles χ2 and ψ± decay into the dark-matter candidate χ1 with a feeble coupling, and the dark-matter yield is set by ΓA/H at temperatures near T ∼ mA/3. The search strategy rests on the decay-length formula cτχ2 ∝ m1/m2² and on the probability Pdecay = exp(−La/βcτ) − exp(−Lb/βcτ) that a boosted parent decays inside a given detector volume; MATHUSLA’s projected reach is computed from this probability times its geometric acceptance, with the 95% C.L. sensitivity set at three signal events in a background-free detector. The relic-density calculation is carried out with a modified version of micrOMEGAS5.0 that includes the thermal masses ΠH(T) and ΠΨ(T) and the electroweak phase-transition temperature TEW.","core_discovery":"In the singlet-doublet freeze-in model studied here, the dark-matter candidate χ1 is produced by the decays χ2 → h χ1, χ2 → Z χ1, and ψ± → W± χ1; requiring ΩDM h² = 0.12 fixes the parent decay length to roughly cτ ≈ 4 km × (m1/100 MeV)(500 GeV/m2)² in the limit m2 ≫ m1, so cosmologically viable parameters naturally have macroscopic lifetimes. The paper shows that this lifetime window overlaps the design sensitivity of MATHUSLA, that combined with the ATLAS displaced-vertex-plus-missing-energy search the two probes cover decay lengths from about a meter to 10⁷ meters and dark-matter masses from the Lyman-alpha bound up to a few GeV, and that a forward detector at a 100 TeV hadron collider would extend parent-mass reach to about 10 TeV and lifetimes up to the BBN bound for m2 ≲ 600 GeV. It also establishes that including the thermal masses of the Higgs doublet and the parent doublet, together with the temperature of the electroweak phase transition, changes the predicted relic density by up to a factor of about five in parts of the parameter space.","pith_inferences":["If the background-free, perfect-efficiency assumption is relaxed, the same approach applied to other freeze-in parents, such as dark photons or heavy neutral leptons, would likely preserve the qualitative complementarity between a surface detector and displaced-vertex searches, but the reach would shrink roughly as the square root of the required event count.","The thermal-mass correction identified here is generic to freeze-in from decay, so existing relic-density calculations for other long-lived-parent models may need a similar revision; the paper states this generality but does not quantify it for other models.","The forward detector’s geometric acceptance of about 0.5 at 100 TeV suggests that even a moderate detector volume placed down the beam line could outperform a large surface detector for very long lifetimes, which could motivate optimising forward geometry in future collider designs.","The assumption that the visible decay products hit the tracking layers whenever the parent trajectory does is only justified for m2 ≫ mh + m1; for compressed spectra near m2 ≈ mh + m1, a full detector simulation could either validate or shrink the claimed sensitivity contours."],"forward_implications":["MATHUSLA100 and MATHUSLA200 could probe dark-matter masses from the Lyman-alpha bound up to m1 ∼ 1–10 GeV, covering a wide region of the viable freeze-in parameter space between the Lyman-alpha and BBN constraints.","For decay lengths cτχ2 ≲ 100 m, the ATLAS displaced-vertex-plus-missing-energy search is the most sensitive probe, while MATHUSLA dominates at larger decay lengths; together they cover decay lengths from about a meter to 10⁷ meters.","A forward detector at a 100 TeV proton-proton collider with 3 or 30 ab⁻¹ of integrated luminosity could probe parent masses up to about 10 TeV and reach lifetimes up to the BBN bound for m2 ≲ 600 GeV.","Including thermal masses and the electroweak phase transition changes the predicted relic density by up to a factor of about five, so earlier freeze-in-from-decay relic computations that omitted these effects can be off by that amount in parts of the parameter space.","The super-WIMP contribution becomes sizable only for m2 > 1.1 TeV and m1 ≳ 300 GeV, and in that region the freeze-in and super-WIMP contributions must be added together when matching the observed dark-matter abundance."],"supporting_citations":[{"why":"Establishes the freeze-in-from-decay production mechanism and the Boltzmann-equation framework used throughout.","marker":"[8]"},{"why":"Defines the singlet-doublet freeze-in model and the LHC displaced-signature analysis that this work extends to MATHUSLA and FCC-hh.","marker":"[30]"},{"why":"Introduces the surface-detector concept for long-lived particles and motivates the forward-detector geometry at a future 100 TeV collider.","marker":"[32]"},{"why":"Provides the MATHUSLA100 and MATHUSLA200 detector volumes and the background-free operating assumption used for the sensitivity projections.","marker":"[34]"},{"why":"Supplies the ATLAS displaced-vertex-plus-missing-energy search whose 13 TeV limit and recast define the LHC sensitivity.","marker":"[42]"},{"why":"Provides the NLO+NLL Higgsino pair-production cross sections used to normalise signal rates at 13 and 100 TeV.","marker":"[44]"},{"why":"Provides the freeze-in relic-density computation that the paper modifies to include thermal masses and the electroweak phase transition.","marker":"[46]"},{"why":"Sets the Big-Bang Nucleosynthesis bound on long-lived decaying particles that defines the upper lifetime limit in the parameter-space plots.","marker":"[65]"},{"why":"Sets the Lyman-alpha lower bound on dark-matter mass for freeze-in produced by two-body decays of a thermal parent.","marker":"[69]"}],"fun_headline_variants":["MATHUSLA and HL-LHC scan freeze-in dark matter parameter space","Long-lived particle searches open freeze-in dark matter to colliders","Freeze-in dark matter from Higgs: MATHUSLA on the trail","Collider long-lived signatures close in on freeze-in dark matter","FCC-hh forward detector extends freeze-in dark matter reach"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The projections assume MATHUSLA and the FCC-hh forward detector are background-free with perfect detection efficiency, so the 95% C.L. limit is set at three signal events, and that the visible χ2 decay products always hit the tracking layers whenever the parent trajectory does; if backgrounds, reconstruction efficiencies, or that acceptance assumption are worse, the projected reach shrinks.","fun_headline_variants_meta":{"raw":{"variants":["MATHUSLA and HL-LHC scan freeze-in dark matter parameter space","Long-lived particle searches open freeze-in dark matter to colliders","Freeze-in dark matter from Higgs: MATHUSLA on the trail","Collider long-lived signatures close in on freeze-in dark matter","FCC-hh forward detector extends freeze-in dark matter reach"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000407,"raw_usage":{"total_tokens":2164,"prompt_tokens":1043,"completion_tokens":1121,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":659,"completion_tokens_details":{"reasoning_tokens":1030}},"tokens_in":659,"tokens_out":1121,"duration_ms":10075,"temperature":1.0,"reasoning_tokens":1030,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:16:31.971209+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A full detector-level simulation of MATHUSLA including cosmic-ray and beam backgrounds, applied to the χ2 → h χ1 signal, would settle the reach: if backgrounds exceed about one event in the signal region at 300 fb⁻¹, the three-signal-event sensitivity contours in Figures 6 and 9 do not hold, and simulating decays with m2 close to mh + m1 would test whether the acceptance assumption that decay products hit the tracking layers holds.","supporting_citations":[{"cited_title":"A Letter of Intent for MATHUSLA: a dedicated displaced vertex detector above ATLAS or CMS","cited_arxiv_id":"1811.00927","evidence_quote":"Provides the MATHUSLA100 and MATHUSLA200 detector volumes and the background-free operating assumption used for the sensitivity projections."}],"review_version":1}