{"id":"e485a2a1-88e3-4e7d-a980-0e950563418e","arxiv_id":"2412.10141","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A parametrized simulation of the IDEA detector at FCC-ee shows that a search for displaced vertices in Zh events can probe Higgs decays to long-lived dark scalars down to a branching ratio of about 10^-4.","lead":"FCC-ee, a proposed electron-positron collider, could detect Higgs bosons decaying into pairs of long-lived dark scalars via displaced vertices. A simulation with the IDEA detector concept suggests sensitivity to branching ratios down to 10^-4 for scalar lifetimes around one meter.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Delphes-based displaced vertex efficiencies drive the BR~1e-4 reach; a full Geant4 validation of LCFIPlus on the IDEA geometry is needed before the projection is treated as quantitative.","rationale":"The reader correctly identifies the parametrized detector simulation as the weakest assumption. I agree: the signal efficiency is the primary multiplier between the physics model and the projected event counts, and the 'zero background' claim is also conditioned on the DV fake rate. The paper's own caveat about background MC statistics is honest but does not remove the need to validate the signal-side reconstruction. The three-event discovery threshold is a secondary issue; even if one treats it as a sensitivity estimate rather than a discovery, the event-count scale factor is still set by the DV reconstruction efficiency. I therefore recommend no change to the reader's CONDITIONAL verdict: the study is a valid and clearly documented sensitivity estimate, but its headline reach at BR~1e-4 should be regarded as conditional on a full simulation cross-check of LCFIPlus on the IDEA geometry.","tokens_in":10844,"tokens_out":6111,"duration_ms":63501,"concrete_test":"Run a Geant4-based full simulation (e.g., the FCCSW/key4hep IDEA model) for one high-sensitivity signal sample (ms=50 GeV, sin theta=3e-7, c tau~1.2 m) and for the dominant Zh background with H->bb, using the same LCFIPlus secondary-vertex configuration and track selections as Secs. 4.3 and 5. Compare acceptance x efficiency and the number of events passing the final selection against the Delphes results in Table 3. If the signal efficiency changes by more than 20% relative, or if the background sample yields at least one event in a 10.8 ab^-1 equivalent, the quantitative BR~1e-4 projection is not robust and should be re-derived with corrected efficiencies. Also report the radial distribution of reconstructed DVs to confirm that vertexing beyond the vertex detector (r>34 cm) is not overestimated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim--sensitivity to BR(h->ss) around 10^-4 for c tau ~1 m--rests on the signal acceptance times efficiency reported in Section 6, which is computed with a parametrized Delphes simulation of the IDEA detector. The custom DV reconstruction (Sec. 4.3) requires |d0|>2 mm seed tracks, at least 3 tracks per vertex, and m_ch>2 GeV, and reconstructs DVs only within r<2 m. The efficiency of this LCFIPlus-based algorithm is sensitive to the track impact-parameter resolution, the material budget, and pattern-recognition losses, none of which are modeled by Delphes with full fidelity. In particular, decays occurring at r>34 cm are inside the drift chamber volume rather than the vertex detector; the parametrization may overestimate the ability to seed and fit vertices from tracks that originate beyond the innermost layers. A modest factor-of-two change in signal efficiency alters the 10^-4 reach materially, and the claim of zero SM background also depends on the rate of fake DVs, which the fast simulation likely underestimates. The paper itself notes the need for larger background MC statistics (Sec. 5), but does not validate the signal-side vertexing efficiency against a full simulation. This is a validation gap, not an internal inconsistency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a fast-simulation sensitivity study for long-lived dark scalars produced in exotic Higgs decays at the proposed FCC-ee collider. Signal events are Zh with Z->ll and h->ss, with each s decaying to b bbar; the final state is selected by requiring a Z-like lepton pair and at least two displaced vertices reconstructed with the IDEA detector concept in Delphes. The analysis uses the FCCAnalyses framework and LCFIPlus-based secondary vertex finding with a custom track selection (|d0|>2 mm, N_trk>=3, m_ch>2 GeV). The authors report zero Standard Model background after the final selection and estimate sensitivity to BR(h->ss) as low as about 1e-4 for c tau around 1 m, based on rescaling a representative signal point from kappa=0.0007 to kappa=0.0002 and requiring three selected events.","tokens_in":11111,"tokens_out":6738,"duration_ms":72060,"significance":"The paper is a well-documented and internally consistent sensitivity projection for a relevant LLP signature at FCC-ee, using the common FCC software stack. Its strengths include the public simulation pipeline, clear event-count tables at each selection stage, and a straightforward branching-ratio rescaling argument. If the projected reach survives the validation concerns below, the result would be a useful benchmark for the FCC-ee physics program and for future detector-concept studies. The analysis also provides concrete insight into the role of track-impact-parameter and vertexing performance for displaced vertices in the IDEA tracker.","major_comments":[{"comment":"The sentence \"The event selection reduces the background to zero and discovery is therefore possible for any signal point with at least three selected events\" conflates discovery with exclusion. In a zero-background counting experiment, observing three events corresponds to a 95% confidence-level upper limit on the signal mean (Poisson: 2.996 events), not to a 5-sigma discovery. The shaded sensitivity region in Figure 7 and the abstract's \"probe\" language should be reframed as exclusion sensitivity unless the authors define a discovery criterion with a specified significance and account for the uncertainty on the zero-background assumption. This is load-bearing because the central quantitative claim (BR below 1e-4) is derived from a point with about 3.3 selected events.","section":"Section 6"},{"comment":"The full sensitivity estimate relies on the displaced-vertex reconstruction efficiency computed with the parametrized Delphes simulation. The custom LCFIPlus configuration (|d0|>2 mm seed tracks, N_trk>=3, m_ch>2 GeV, vertices up to r=2 m) is not validated against a Geant4-based simulation of the IDEA detector, and this is particularly relevant for decays occurring inside the drift-chamber volume at r>34 cm, where track seeding and vertex fitting depend on pattern recognition and material effects that Delphes models only parametrically. The reported reach is linearly proportional to this efficiency: for the (50 GeV, 3e-7) sample, a factor-of-two reduction in signal efficiency would lower the rescaled yield from 3.3 to about 1.6 events, dropping below the three-event threshold. I request that the authors either validate the vertexing efficiency with a full simulation, or provide a systematic uncertainty on the efficiency and quote the resulting uncertainty on the BR reach, or explicitly state that the quoted reach is contingent on the unvalidated Delphes vertexing performance.","section":"Section 4.3 and Section 6"},{"comment":"The conclusion that the selection is background-free is based on zero events surviving from approximately 2.54 million pre-selected background events. With zero observed events, the 95% confidence-level upper limit on the background yield is about three events divided by the number of pre-selected events, i.e., roughly 1.2e-6 per pre-selected event if one treats the pre-selection count as fixed. This statistical bound should be reported explicitly, and it should be propagated into the sensitivity statement, because the \"discovery possible\" claim assumes exactly zero background. The paper's note that larger MC statistics would be beneficial is appropriate, but it does not quantify the impact on the projected reach.","section":"Section 5, Table 3"}],"minor_comments":[{"comment":"The sentence \"The study would therefore benefit from larger MC statistics\" would be more useful if accompanied by a quantitative target, such as the number of background events needed to place a given upper limit on the background expectation.","section":"Section 5"},{"comment":"References [10] and [31] appear to share the same DOI (10.17181/nfs96-89q08); please verify that [31] points to the intended vertex-fitting package documentation.","section":"References"},{"comment":"The y-axis label in Figure 2 appears truncated or malformed (\"10^5 10\"); please ensure the axis labels and exponents render correctly.","section":"Figure 2"}],"recommendation":"major_revision","confidential_remarks":"The paper is a credible fast-simulation sensitivity study and appears well-suited to JHEP in scope. The main risk is not internal inconsistency but the quantitative reliance on unvalidated Drift-chamber-era vertexing efficiencies in Delphes; this should be addressed before publication. I do not see any novelty-disclosure or citation-pattern concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a clean, well-documented sensitivity projection, and the first with the IDEA detector concept in the common FCC framework. The central number — sensitivity to BR(h→ss) around 1e-4 for cτ≈1 m at 10.8 ab⁻¹ — is plausible as a planning estimate, not as a sharp discovery claim.\n\nWhat's new: prior LLP work at lepton colliders (Alipour-Fard et al.) used different detector assumptions; this paper gives the first IDEA-specific study, with a full chain from MadGraph/Pythia through Delphes and LCFIPlus in FCCAnalyses. The background samples are sizable (hundreds of millions of events), the event selection is well motivated (Z-tagged dilepton plus ≥2 displaced vertices), and the acceptance tables are useful. The paper is also honest: it notes the need for larger background MC and shows the radial DV distributions, which look physically sensible.\n\nSoft spots, in proportion. The 'discovery' threshold of three events is statistically weak — about 2σ in a zero-background counting experiment. For a projection this is common, but the shaded reach region should not be read as a real 5σ claim. The zero-background statement itself is based on finite MC; the 33M Zh / 56M ZZ / 373M WW events give good coverage, but after all cuts zero events doesn't mean the background is exactly zero. More importantly, the signal-side vertexing efficiency comes from a parametrized Delphes simulation of the IDEA tracker. The cuts (|d0|>2 mm seeds, ≥3 tracks, m_ch>2 GeV, DV radius <2 m) are exactly where pattern recognition and impact-parameter resolution matter most, and Delphes cannot model those with full fidelity. A factor-of-two change in signal efficiency would shift the 1e-4 reach noticeably. The paper flags its own limitation on the background side but does not validate the signal-side vertexing against Geant4. That's a validation gap, not an internal inconsistency; I don't see circularity or anything contrived in the selection.\n\nBottom line: for FCC-ee physics-case people and LLP hunters, this is a useful, citable study. It deserves a serious referee — the analysis is reproducible in structure, the limitations are mostly stated, and a referee's main job would be to push on the DV-efficiency question and the statistical definition of reach. I'd be happy to see it published after those points are either addressed or clearly caveated. Recommend: send to peer review.","headline":"A clean, internally consistent FCC-ee sensitivity projection for displaced dark scalar decays; the reach numbers are plausible but rest on a fast-simulation vertexing efficiency that needs a full Geant4 check before being treated as quantitative.","tokens_in":11657,"tokens_out":2406,"would_cite":true,"duration_ms":26431,"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":"A background-free selection based on a Z-like lepton pair plus two displaced vertices can probe Higgs decays to long-lived dark scalars at branching ratios down to $10^{-4}$ for lifetimes near 1 m.","keywords":["long-lived dark scalars","exotic Higgs decays","displaced vertices","FCC-ee","IDEA detector concept","Higgs portal","ZH production","background-free search"],"falsifier":"Run a full detector simulation of $e^+e^-\\to Zh\\to l^+l^-b\\bar{b}b\\bar{b}$ for the 50 GeV, $3\\times10^{-7}$ signal point and for the dominant $Zh$, $ZZ$, and $WW$ backgrounds, and count displaced vertices satisfying $N_{\\rm trk}\\ge3$ and $m_{\\rm ch}>2$ GeV; if the reconstructed-vertex rate or the zero-background result changes by more than the statistical uncertainty, the claimed $10^{-4}$ branching-ratio reach would need revision.","tokens_in":10641,"feed_emoji":"⚛️","tokens_out":8598,"duration_ms":79056,"temperature":0.7,"pith_summary":"This paper claims that a dedicated search for displaced vertices at the FCC-ee can probe exotic Higgs decays into pairs of long-lived dark scalars with branching ratios down to $10^{-4}$, covering dark scalar masses from 20 to 60 GeV and mean proper lifetimes $c\\tau$ from roughly 10 mm to 10 m. The signal is $Zh$ production at $\\sqrt{s}=240$ GeV with the $Z$ decaying leptonically and the Higgs decaying to two dark scalars that each decay to $b\\bar{b}$, so the final state contains at least two displaced vertices. Using a fast, parametrized simulation of the IDEA detector concept and an event selection that requires a Z-like lepton pair plus two or more displaced vertices, the authors find zero Standard Model background in the simulated samples. This is the first sensitivity estimate for exotic Higgs decays at FCC-ee with the IDEA detector concept, and it suggests a discovery-level channel complementary to LHC searches.","feed_headline":"FCC-ee could probe Higgs-to-dark-scalar decays down to 1e-4","feed_subtitle":"Z pair + two displaced vertices leaves zero Standard Model background, opening a window on long-lived dark scalars.","key_machinery":"The load-bearing mechanism is the displaced-vertex selection built on non-primary tracks. Tracks with $p_T>1$ GeV and $|d_0|>2$ mm seed a secondary-vertex finder; vertices are required to have at least three tracks, a charged invariant mass above 2 GeV, and to lie within the 2 m radial tracker volume. The $|d_0|$ cut is what suppresses heavy-flavour background, the mass and multiplicity cuts reject random track crossings, and the final requirement of at least two such vertices, together with a $Z$-mass lepton pair, eliminates all simulated Standard Model processes.","core_discovery":"The central claim is that a background-free search for $h \\to ss \\to b\\bar{b}b\\bar{b}$ is achievable at the FCC-ee. The selection starts from events with exactly two opposite-charge, same-flavour leptons whose invariant mass lies between 70 and 110 GeV, then requires at least two reconstructed displaced vertices satisfying track multiplicity $N_{\\rm trk}\\ge 3$ and charged invariant mass $m_{\\rm ch}>2$ GeV. With this selection, all $Zh$, $ZZ$, and $WW$ background events in the simulation are removed, while a grid of signal samples with $m_s$ between 20 and 60 GeV and $\\sin\\theta$ between $10^{-7}$ and $10^{-5}$ retains up to 32% acceptance. The authors argue that rescaling to lower coupling would allow probing $\\text{BR}(h\\to ss)$ below $10^{-4}$ at $c\\tau\\approx 1$ m, with the best point (50 GeV, $3\\times10^{-7}$) giving 3.3 selected events at a branching ratio of $9\\times10^{-5}$.","pith_inferences":["If the background-free behaviour persists with larger Monte Carlo statistics and a full detector simulation, the same two-displaced-vertex selection could be applied to other long-lived decays from exotic Higgs decays, such as charm or tau final states, by retuning the vertex mass and track multiplicity cuts.","The 2 m radial reach of the tracker sets a natural upper limit on the lifetime probed; adding timing information or an outer decay volume would extend sensitivity beyond $c\\tau\\sim10$ m.","Combining the ZH run with other Higgs production modes at FCC-ee, such as WW fusion, could improve the branching-ratio reach or cover additional mass-lifetime combinations beyond the grid studied here."],"forward_implications":["At $\\kappa=0.0002$, the 50 GeV, $3\\times10^{-7}$ sample corresponds to $\\text{BR}(h\\to ss)=9\\times10^{-5}$ and gives about 3.3 selected events, so the search could realistically probe branching ratios below $10^{-4}$.","Zero background in simulation means discovery would be possible for any signal point with at least three selected events.","The selection covers $m_s$ from 20 to 60 GeV and $c\\tau$ from about 10 mm to 10 m; the edges lose sensitivity because decays occur inside 2 mm or beyond 2 m.","The 50 GeV, $3\\times10^{-7}$ point has the highest acceptance, about 32%, since 99% of its dark scalars decay inside the reconstructible tracker volume."],"supporting_citations":[{"why":"defines the exotic Higgs decay framework and the scalar mixing model, including the dominant $s\\to b\\bar{b}$ branching ratio used for normalization","marker":"[1]"},{"why":"established that FCC-ee sensitivity to long-lived particles from exotic Higgs decays can compete with the HL-LHC, motivating this search","marker":"[6]"},{"why":"provides the FCC-ee ZH run integrated luminosity of 10.8 ab$^{-1}$ and the 2.2 million expected $Zh$ events used for normalization","marker":"[10]"},{"why":"describes the IDEA detector concept whose tracking geometry and performance define the signal acceptance","marker":"[11]"},{"why":"supplies the HL-LHC displaced-vertex projection below $10^{-4}$ that the FCC-ee reach is compared with","marker":"[14]"},{"why":"provides the Hidden Abelian Higgs Model implementation used to generate $h\\to ss\\to b\\bar{b}b\\bar{b}$ events","marker":"[15]"},{"why":"provides the fast parametrized detector simulation used to reconstruct tracks, leptons, and vertices","marker":"[27]"},{"why":"supplies the secondary vertex finder algorithm that is reconfigured with custom track selections for displaced vertices","marker":"[32]"}],"fun_headline_variants":["FCC-ee sees zero background for long-lived dark scalars","Dark scalar search at FCC-ee probes branching ratios down to 1e-4","Background-free window on exotic Higgs decays at FCC-ee","Hunting long-lived dark scalars with displaced vertices at FCC-ee","FCC-ee could unmask dark scalars with zero Standard Model noise"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The search assumes the fast, parametrized simulation of the IDEA detector reconstructs displaced vertices at radii up to two metres with realistic efficiency and resolution, a point that is not checked against a full detector simulation.","fun_headline_variants_meta":{"raw":{"variants":["FCC-ee sees zero background for long-lived dark scalars","Dark scalar search at FCC-ee probes branching ratios down to 1e-4","Background-free window on exotic Higgs decays at FCC-ee","Hunting long-lived dark scalars with displaced vertices at FCC-ee","FCC-ee could unmask dark scalars with zero Standard Model noise"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000208,"raw_usage":{"total_tokens":1471,"prompt_tokens":1079,"completion_tokens":392,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":695,"completion_tokens_details":{"reasoning_tokens":296}},"tokens_in":695,"tokens_out":392,"duration_ms":4982,"temperature":1.0,"reasoning_tokens":296,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:17:05.914068+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a full detector simulation of $e^+e^-\\to Zh\\to l^+l^-b\\bar{b}b\\bar{b}$ for the 50 GeV, $3\\times10^{-7}$ signal point and for the dominant $Zh$, $ZZ$, and $WW$ backgrounds, and count displaced vertices satisfying $N_{\\rm trk}\\ge3$ and $m_{\\rm ch}>2$ GeV; if the reconstructed-vertex rate or the zero-background result changes by more than the statistical uncertainty, the claimed $10^{-4}$ branching-ratio reach would need revision.","supporting_citations":[{"cited_title":"C 43 (2020) 27","cited_arxiv_id":null,"evidence_quote":"describes the IDEA detector concept whose tracking geometry and performance define the signal acceptance"}],"review_version":1}