Search for long-lived, heavy particles in final states with a muon and multi-track displaced vertex in proton-proton collisions at sqrt(s)=7 TeV with the ATLAS detector

Nothing here will surprise you, and that is fine. This is a competent displaced-vertex search with zero events in the signal region, an expected background of about 0.004, and a 95% CL limit of 0.68 fb on sigma times BR squared times acceptance times efficiency. The arithmetic checks out: 3.0/4.4 fb⁻¹ is exactly the Poisson limit for zero events with negligible background, and the paper says so plainly rather than dressing it up. The paper is also candid about its own costs: Section 9 admits that for cτ < 100 mm the limits are weaker than naive luminosity scaling would suggest, because the trigger and offline muon selections were tightened relative to the 2010 analysis. That kind of statement is worth more than a page of hedging elsewhere. Two things are genuinely new and worth crediting. First, the re-run of the silicon-seeded tracking with looser impact-parameter and shared-hit requirements (Section 4, Fig. 2), which recovers a large fraction of displaced tracks beyond about 10 mm. Second, requiring the muon to be associated with the displaced vertex, which makes the per-neutralino selection efficiency factorizable and lets the limit be reinterpreted for events with different numbers of long-lived particles. That is a real design choice, not a cosmetic one. The soft spots are the ones your reader already named, and they are small. The efficiency comes from three simulated mass points mapped onto a two-dimensional grid and reweighted in cτ, with no data cross-check at large lifetime. The hadronic-interaction background is validated only in the mDV < 10 GeV sideband, so the extrapolation into the signal region is assumed rather than demonstrated. Neither gap threatens the qualitative claim, since the background central value is 0.004 and the exclusion survives a factor-of-two error in efficiency. One presentational point the stress-test raised is worth passing to the authors: Fig. 4 shows the efficiency map only for the MH sample, and the text never states whether ML and HH use their own maps. Fig. 3 implies they do, but a sentence would settle it. This is a paper for people building long-lived-particle models or writing reinterpretation tools, not a general-interest result. It deserves referees, and I expect it to pass with minor comments.

The limit on σ·BR²·A·ε is 0.68 fb = 3.0/4.4 fb⁻¹, exactly the Poisson 95% upper limit for zero observed events with ~0.004 expected background, so the arithmetic is internally consistent. To convert to σ·BR² (Fig. 6), the paper divides by ε(cτ) from Fig. 3, which is derived from three MC mass points and a 2D (|z_DV|, r_DV) map reweighted in cτ. The limit scales as 1/ε, so a bias in ε directly biases the quoted limit. The paper validates secondary-track reconstruction with K_S (Section 7) and checks the HH-vs-MH boost reweighting (Section 5), but does not directly validate the full vertex-finding efficiency at large cτ with signal-like data. Additionally, Fig. 4 shows the map only for MH; the text does not explicitly state whether ML and HH use their own maps or the MH map, which matters because track multiplicity and pT depend on neutralino mass. However, the systematic uncertainties (muon ~7%, trigger ~6%, tracking ~4%) are folded into the CLs limit, and the qualitative claim (zero events, 700 GeV excluded) is robust to factor-2 changes in ε. So the concern is real but not decisive.

The paper's central claim is a null result translated into a limit. The calculation is internally consistent: 0.68 fb = 3.0/4.4 fb⁻¹, matching the Poisson 95% upper limit for zero observed with negligible background. The background is data-driven and tiny, so it is not load-bearing. The efficiency is the dominant input, and the reader correctly identifies it as the weakest assumption. However, the paper does more validation than the reader credits: the K_S study cross-checks secondary-track reconstruction over a range of radii, and the HH-vs-MH boost reweighting checks the kinematic reweighting. The residual concern is the full vertex-finding efficiency at large cτ, which is not directly validated with signal-like data, and the ambiguity about whether the 2D map is per-sample. These are real but not decisive: the systematic uncertainties are folded into the limit, and the main exclusion (700 GeV) is robust to factor-2 changes in ε. I also note that the reader's strongest_claim overstates the paper's reach by including 1500 GeV in the 'excludes' list; based on the numbers (PROSPINO 0.2 fb vs limit ~2 fb), 1500 GeV is not excluded, and the paper correctly says 'limits are set' rather than 'excluded.' This is a reader paraphrase issue, not a paper error. Overall, the paper is solid and the verdict should remain ACCEPT.

This is a competent, well-motivated search that improves substantially on the earlier ATLAS result [16] by using a dataset roughly 100 times larger, by recovering secondary tracks through a dedicated re-tracking pass, and by removing the dependence on simulation for the background estimate. The data-driven background extraction in Section 6, built from jet-triggered control samples, low-mass vertex sidebands, and material-layer versus air-gap vertex ratios, is a genuine strength, as is the explicit factorization of the per-neutralino efficiency through the muon-DV association requirement, which is what makes the limits reusable in models with different numbers of long-lived particles per event. The systematics are enumerated with sizes (muon about 7%, trigger about 6%, tracking about 4%, luminosity 3.9%), and the CLs machinery is standard. If the efficiency mapping holds, the resulting exclusion is a useful and quotable constraint on RPV SUSY with displaced neutralino decays.

Many proposed extensions of the Standard Model contain heavy particles that do not decay immediately. A particle with a lifetime of picoseconds to a nanosecond travels a few millimetres to tens of centimetres before decaying, so its decay products form a vertex visibly separated from the proton collision point. Finding such a vertex would be direct evidence of new physics. This paper looks for that signature in 4.4 inverse femtobarns of 7 TeV proton-proton collisions recorded by ATLAS in 2011. The trigger requires a muon with transverse momentum above 40 GeV. Offline, the analysis demands a muon above 50 GeV that misses the primary vertex by more than 1.5 mm, then searches the inner detector for a vertex made of at least five tracks, at least 4 mm from any primary vertex, with an invariant mass above 10 GeV, and requires the muon to point back to it. Regions of dense detector material are masked out because hadronic interactions there fake displaced vertices. The dominant technical work is estimating the background without relying on simulation. Random crossings of tracks and real hadronic interactions in gas are extrapolated from control regions in data, giving an expected background of about 0.004 events. No events survive. The result is a 95% confidence-level upper limit of 0.68 fb on the production cross section times the square of the branching fraction, for a benchmark R-parity-violating supersymmetry model, plotted against the neutralino lifetime and for several squark and neutralino masses.

Core claim. From the abstract and Section 8: 'No events are observed in the signal region' and 'a 95% confidence-level upper limit of 0.68 fb is set on sigma times BR squared multiplied by the detector acceptance and the reconstruction efficiency for any signal vertex', in 4.4 fb^-1 of 7 TeV data. If correct, the paper excludes, at 95% CL, R-parity-violating supersymmetry scenarios in which pair-produced squarks give neutralinos that decay to a muon plus quarks at displaced vertices, over a range of neutralino lifetimes and for squark masses of 700 and 1500 GeV.

The premise it rests on. The limit scales inversely with the signal efficiency, and that efficiency is taken entirely from three simulated mass points (Table 1) mapped onto a two-dimensional (|z_DV|, r_DV) grid (Fig. 4) and then reweighted in cτ. The premise, stated in Section 5, is that this map plus a decay-position reweighting describes acceptance for arbitrary squark and neutralino masses and lifetimes. Nothing in the data cross-checks the efficiency at large cτ, where the re-tracking gain (Fig. 2) is largest and least constrained; the HH-versus-MH boost reweighting check covers kinematics but not the tracking model itself.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review