Constraining New Physics Models after Three Years of LHC Operation and Preparation of SUSY Searches During the Second LHC Run
Final Report Abstract
I worked on two different, though related projects during my time at CBPF. The first project was the preparation of a global fit of the pMSSM. While it was originally planned to already perform the fit during the funding period, a lot of work had to be put into the further developement of the Fittino software framework. This was necessary to achieve the desirable performance of the software during the fit, as well as reliable physics results. In particular, a method was developed to accurately determine the p-vlaue of the pMSSM using pseudo experiments, without having to rerun the full fit for each set of pseudo measurements. The method was tested with toy models and proved to give reliable results. Applied to a fit of the CMSSM, the p-value of that model was determined for the first time with pseudo experiments. The obtained χ2 -distribution differed significantly from the naive expectation assuming all gaussian observables and a linear model, giving a p-value of 4.9% for the CMSSM with the baseline observable set, excluding the CMSSM at the 90% CL. For the fit of the pMSSM with 11 free SUSY breaking parameters, the Fittino software was extended, improved and tested. In particular, an effective sampling algorithm for the complex and highly correlated parameter space in the pMSSM was developed and tested and the consideration of the most interesting final states analysed by the ATLAS and CMS experiments was prepared by the Fittino collaboration and is currently being tested. The second project was the preparation of a direct collider search for SUSY in a compressed scenario with the CMS detector at the LHC. A simplified model with nearly mass-degenerate gluinos and neutralinos and heavy squarks was derived from the pMSSM motivated by preliminary results of a pMSSM prefit, providing a realistic image of the pMSSM with this particular mass hierarchy. In this scenario, gluinos can be produced at the LHC with a high cross-section. The gluinos are metastable, form R-hadrons and decay at an average distance of up to several 10mm from the primary interaction point into two low energetic quarks and the lightest neutralino. The signature in this case are two jets originating from a reconstructed secondary vertex, well separated from the interaction point, as well as a certain amount of missing transverse energy. In order to be able to select such events in real data taking, a high energetic ISR jet is required in order effectively trigger on these events. The standard CMS reconstruction algorithms were tested and found to provide a sufficiently high selection efficiency with a custom jet-vertex association algorithm for the signal. A preliminary signal region was defined, providing an expected sensitivity for gluino masses , with O(1) expected background events from SM processes. Techniques for a datadriven estimation of the two major backgrounds, QCD multijet production and top pair production, were developed - these methods are still under development. For the first iteration of the analysis, the background will therefore likely be estimated from SM MC. The modelling of the background in the MC was validated using the full 2015 dataset and well defined control regions. The first iteration of the analysis is currently being prepared for preapproval within the CMS SUSY group.
Publications
-
Eur. Phys. J. C76(2), 96 (2016)
P. Bechtle, et al.
