CMSDAS: Muon Detector Shower (MDS) Long Exercise: Useful links

Searching for LLPs with the CMS muon detectors that is interleaved with steel return yoke give rise to unique high multiplicity signature that allow us to be sensitive to a broad range of LLP decay modes and to LLP masses below GeV

Particle proper decay lengths follow an exponential decay.

Long-lived particles for LHC searches generally have mean proper decay lengths of ~mm to ~km scale that create displaced signatures in the sub-detectors

LLP lifetime can be reweighted to avoid generating many signal samples with different LLP lifetimes

The probability that an LLP decays in a sub-detector depends on the LLP mean proper decay lengths, so searches for LLPs with different sub-detectors (tracker, calorimeter, and muon detectors) provide complementary coverage to LLPs

MDS is a cluster of rechits in the muon system, clustered by the DBSCAN algorithm

MDS properties are computed from the input rechits(e.g. position,time & station) and are very powerful of rejecting background

MDS reconstruction efficiency depends on where the LLP decays with respect to the steel, since the decay particles require small amount of steel to initiate the shower and are detected only in the active gas chambers

Due to the lack of dedicated trigger, we use the high MET trigger in Run 2 to trigger on the signal

The background from main collision comes from punch-through jet and muon bremsstrahlung and are killed by dedicated jet and muon vetos and active vetos using the first muon detector station

The remaining irreducible background comes from low pT particles from pileup events and clear out-of-time pileup contributions can be observed from cluster time distribution

Two final discriminating variables that are independent for background will be used to extract signal and estimate background

The ABCD method requires the use of two independent variables for background, which implies the only source of background should be low pT particles

The method has been validated in the out-of-time validation region, allowing us to proceed to statistical analysis in the signal region

We can use Higgs Combine to perform statistical analysis and set limits

The value of the limits depend significantly on the LLP mean proper decay length, as the probability of LLP decaying in the muon system strongly correlates with the LLP lifetimes

Event displays allow us to view all available collections visually and scrutinize event topologies that are not possible with ntuples