| One of the largest mysteries in astronomy is that a vast amount of "dark matter"---more than four times as much as ordinary matter---is required to explain several observations and yet remain within Newtonian laws of gravitation. If dark matter does in fact exist, it should be possible to produce it at a high-energy collider experiment like the Large Hadron Collider (LHC) experiment. This document describes such a search utilizing 36pb -1 integrated luminosity of LHC data collected by the Compact Muon Solenoid (CMS) detector, with fully hadronic jets plus missing transverse momentum signature. The search regions are designed to be simple, with high efficiencies for a wide variety of Beyond Standard Model signals that have dark matter candidate particles in the event final state. Standard Model processes with high cross-sections like top quark pair-production, W and Z boson production in conjunction with jets, as well as multi-jet production, all form significant backgrounds to these search regions, and are estimated using data-driven techniques. In particular, a novel method for predicting the multi-jet background has been developed, which is by construction highly safe against contamination from other processes including any signals that may be present. This method, called Rebalance and Smear, first unfolds events in an inclusive data sample back to a well-balanced, particle-level description of the jets per event, then smears the momenta of these jets to simulate detector-related measurement effects as would cause multi-jet events to contribute to the search regions. The collider data sample collected in 2010 is seen to be compatible to within one standard deviation of the predicted background yields in all search regions. Limits are set in some slices of the Constrained Minimal Supersymmetric Standard Model parameter space, which extend to gluino and squark masses of 700GeV in some regions and gluino masses nearly up to 500GeV for more or less arbitrary squark masses. The results are also interpreted in terms of Simplified Model Spectra signals, which furthermore provide an intuitive language in which strengths and weaknesses of the search are diagnosed. |
