• analysis
• associated
• ATLAS
• CP
• GBR
• Hbb
• HEP
• LHC
• multi-dimensional
• multivariate
• production
• reweighter
• ZH

The discovery of the Higgs boson in 2012 by the ATLAS and the CMS collaborations at the Large Hadron Collider was pivotal in particle physics. The existence of the Higgs boson, predicted in 1964, is deeply related to the mechanism of electroweak symmetry breaking, and the value of its mass, in conjunction with the value of the mass of the top quark, indicates whether the present electroweak vacuum state
of the universe is likely to be stable or not.
Given the importance of this new boson, its discovery has been followed by an extensive campaign of measurements of its properties. The analyses of the data collected at sqrt(s) = 7, 8 TeV during the LHC Run1 and at sqrt(s) = 13 TeV during the LHC Run 2 led the ATLAS and the CMS collaborations to observe the leading production modes (gluon gluon fusion, vector boson fusion, production in association with a vector bosons, and production in association with a top quarks), and to observe several decay modes (photon-photon , ZZ, W-W+, tau-tau, bb). By using these datasets, the ATLAS and the CMS collaborations measured the mass of the Higgs boson (mH = 124.97 +/- 0.27 GeV in the latest measurement), its couplings, its spin and its parity, with no significant deviation from the prediction of the Standard Model.
The latest observations, in summer 2018, are the production in association with a vector boson (pp -> VH – where V can be either a W or a Z boson) and the decay mode in b-quarks (H -> bb), both guided by the analysis of the pp -> V H(bb) channel.
This thesis investigates the possibility to obtain information on the CP properties of the Higgs boson exploiting this channel. The scope is to develop kinematic variables which are sensitive to the CP structure of the HVV coupling.
This thesis profits of the analysis developed by the ATLAS collaboration that led to the evidence and observation of pp -> VH and of H -> bb.
This work re-uses the event selection of the ATLAS analysis, and it contains two major personal contributions: i) the development of original and alternative tools to model the systematic uncertainties on the backgrounds, which strongly affect this kind of analyses, and ii) the study of
the CP structure of the HVV coupling using the pp -> ZH -> llbb channel.
In the current pp -> VH(bb) analysis, the signal over background ratio is approximately 2% after a typical event pre-selection. An optimization on the sensitivity is reached thanks to multi variate techniques. As a consequence, any future measurement of the pp -> VH process in the H -> bb channel will depend on the capabilities of the experimental collaborations to model with high accuracy the backgrounds.
The estimate of the background is based on Monte Carlo simulations, but their reliability has been verified by only few measurements in the past.
The program of the experimental collaborations foresees new accurate measurements of these backgrounds and improvements in the models used in simulations. An additional important aspect is how to include
these models and their variations into the pp -> VH(bb) analysis. A natural procedure would be to generate Monte Carlo simulations for all possible variations of a certain background model, and include them in the appropriate statistical treatment of the analysis. The weakness of this option comes from the fact that the simulation of all the needed samples asks for very high computing resources (CPU and disk space). An attractive alternative is to change the shape of the multidimensional distribution
of the nominal background model to match the one of an alternative model. This can be done with a multidimensional re-weighting of several kinematic variables, using event-by-event weights. This method
is using only the theoretical predictions and it would prevent from simulating the particles-detector interactions (which typically are the most resource demanding) for the alternative models.
My original contribution has been the testing of multidimensional reweighting techniques which allows to consistently change the shape of several kinematic variables. This is done by re-using, in an original way, the regression techniques based on gradient boosting algorithms, and demonstrating that it is possible to obtain a good re-weighting of O(10) variables, and that it is a suitable option for future analysis.
The second topic, which is the main subject of this work, is to develop variables to study the CP properties of the Higgs boson in pp -> ZH -> llbb. The ATLAS and CMS collaborations already excluded alternative hypothesis for CP++ state of the discovered Higgs boson. On the other side, some models beyond Standard Model, like the Two Higgs Doublet Model, predict the existence of more scalar and pseudoscalar bosons. In these models, if the Higgs potential is not CP symmetric, the lightest mass eigenstate could be a superposition of CP even and odd states, introducing the possibility of non trivial CP structures in the Higgs sector. This option has been already studied by the ATLAS and CMS collaborations, using the kinematic information in the WW and ZZ decays, and more recently,
using the production modes. In this context, the study of pp -> VH production mode was proposed in many theoretical work.
The effects foreseen by the mixing of CP states are encoded in modifications of the HVV coupling structure, with a mixing of the SM couplingwith a CP-odd tensor. Monte Carlo simulations of events produced by these changes have been used to study the performance of the developed
kinematic variables used to investigate the Higgs CP. This thesis represents a first original attempt to study the Higgs CP in this production mode using the data collected by the ATLAS experiment, profiting of the recent observation of the pp -> VH(bb) production.