Studies of Higgs Boson Signals Leading to Multi-Photon Final States with The ATLAS Detector

Abstract

The efficient identification of photons is a crucial aspect in the search for the Higgs boson at ATLAS. With the high luminosity and collision energies provided by the Large Hadron Collider, rejection of backgrounds to photons is of key importance. It is often not feasible to fully simulate background processes that require large numbers of events, due to processing time and disk space constraints. The standard fast simulation program, ATLFAST-I, is able to simulate events ∼1000 times faster than the full simulation program but does not always provide enough detailed information to make accurate background estimates. To bridge the gap, a set of photon reconstruction efficiency parameterisations, for converted and unconverted photons, have been derived from full simulation events and subsequently applied to ATLFAST-I photons. Photon reconstruction efficiencies for isolated photons from fully simulated and ATLFAST-I, plus parameterisations, events are seen to agree within statistical error. A study into a newly proposed Two Higgs Doublet Model channel, gg → H → hh → γγγγ, where the light Higgs (h) boson is fermiophobic, has been investigated. The channel is of particular interest as it exploits the large production cross-section of a heavy Higgs (H) boson via gluon-fusion at the LHC in conjunction with the enhanced branching ratio of a light fermiophobic Higgs (h) boson to a pair of photons. This channel is characterised by a distinct signature of four high pT photons in the final state. Samples of signal events have been generated across the (mh,mH) parameter space along with the dominant backgrounds. An event selection has been developed with the search performed at generator-level. In addition, the search was also performed with simulated ATLFAST-I events utilising the above photon reconstruction efficiency parameterisations. For both analyses, the expected upper limit on the cross-section at 95% confidence level is determined and exclusion regions of the (mh,mH) parameter space are defined for integrated luminosities of 1 f b−1 and 10 f b−1 in seven fermiophobic model benchmarks.