A Global Event Description using Particle Flow with the CMS Detector

The particle-flow reconstruction algorithm aims at providing a global (i.e., complete and unique) event description at the level of individually reconstructed particles, with an optimal combination of the information coming from all CMS subdetectors. The reconstructed and identified individual particle list includes muons, electrons (with individual reconstruction and identification of all Bremsstrahlung photons), photons (either unconverted or converted), charged hadrons (without or with a nuclear interaction in the tracker material), as well as stable and unstable neutral hadrons. These particles can be non isolated, and even originate from a intricate overlap of reconstructed charged particles, ECAL and HCAL energy clusters, and signals in the muon chambers. The particle-flow algorithm starts with reconstruction performed independently within each CMS sub-detector: clustering is conducted separately in each of the electromagnetic (ECAL) and hadronic (HCAL) calorimeters and track reconstruction takes place in the combined silicon and pixel tracker system. The efficiency of tau identification is directly proportional to the tracking efficiency. In addition, the energy of each charged particle missed by the tracking algorithm is determined with the limited hadron calorimeter resolution, and its direction is biassed by the large magnetic field. Conversely, fake tracks may rapidly dominate the jet energy and angular resolutions, and may dramatically increase the QCD background to tau jets. An iterative tracking algorithm was therefore developed in the context of particle flow to improve the overall CMS tracking efficiency and fake rate.The iterative tracking strategy is as follows. First, tracks with a very pure seeding algorithm (such as the requirement of three hits in the pixel detector, the requirement of a minimum number of hits in the silicon tracker, with tight vertex constraints) are reconstructed. This first iteration leads to a moderate efficiency and a very small fake rate. The next steps proceed by cleaning hits used in the previous iteration and by progressively loosening track quality criteria, thus increasing the tracking efficiency without noticeably changing the fake rate. With this technique, charged particles with as little as three hits and a transverse momentum as small as 300 MeV/c can be reconstructed with a larger efficiency and a fake rate similar to those. The original one-step reconstruction which required at least eight hits and a transverse momentum in excess of 0.9 GeV/c Tracks originating from secondary vertices (e.g., from the nuclear interaction of a hadron or a photon conversion in the tracker material, or from the decay of unstable neutral hadrons) are seeded with specific algorithms from the hits remaining after the first-steps cleaning. The particle-flow calorimeter clustering algorithm is conceptually identical for ECAL crystals and HCAL cells. Clusters are seeded by local energy maxima in topologically-connected sets of cells. The energy of any nearby crystal/cell is then shared between individual clusters according to a crystal/cell-to-cluster distance, allowing for better spatial resolution of overlapping clusters and with less biased cluster position measurements than the ECAL clustering algorithms developed, e.g., for isolated electrons and photons. Links between ECAL clusters, HCAL clusters, and tracks are made, depending on the spatial and energy compatibility between the clusters and/or tracks. The particle-flow algorithm proceeds in stages, associating clusters and tracks with a newly reconstructed particle at