Layer Intercalibration of the ATLAS Electromagnetic Calorimeter and CP-odd Higgs Boson Couplings Measurements in the Four-Lepton Decay Channel with Run 2 Data of the LHC

After the Higgs boson discovery at the LHC in 2012, interest turned to Higgs boson property measurements to refine the tests of the Standard Model and probe for new physics. One of its key properties is its spin-parity (CP), predicted to be $0^+$ in the Standard Model. Analyses of data collected during the Run 1 of the LHC rejected all pure spin-parity state other than $0^+$. However mixed CP states are still possible, and would indicate CP violation in the Higgs sector. The first part of this thesis focuses on the ATLAS electromagnetic calorimeter calibration, needed to reach a permil level on electron and photon energy resolution which are of prime importance for Higgs boson studies. One step of the calibration sequence consists of the layer intercalibration of the electromagnetic calorimeter, needed to correct residual electronics miscalibration and cross-talk effects. The Run 1 method has proven to be unreliable for the pileup levels in Run 2 and a new method was developed, ensuring a precise control on the systematic uncertainties. The second part of this thesis puts emphasis on the Higgs boson to vector boson CP-odd couplings, with the Higgs boson decaying to four leptons. This channel, despite low statistics, provides a clean signature and a signal-to-noise ratio over two, allowing for a precise determination of the Higgs boson properties. The vector boson fusion production channel offers the best sensitivity to CP effects thanks to its two characteristic tagging jets in the final state. The contamination from the gluon fusion production mode with additional jets is reduced using neural networks. To unambiguously distinguish yet unknown CP-even from possible CP-odd effects, a variable whose shape asymmetry only depends on CP-odd effects is built. This observable is based on the matrix element computation, maximally using the kinematic information available from Higgs boson and associated jets. Results are interpreted in a context of effective field theory, and the statistical precision on the $\tilde{c}_{zz}$ Wilson coefficient is estimated to $[-0.80, 0.80]$ at the 68% confidence level.