CERN Accelerating science

Thesis
Report number	CERN-THESIS-2021-322 ; CMS-TS-2022-030
Title	Searches for Higgs boson pair production in $b$$\bar{b}$$\gamma$$\gamma$ final states at Compact Muon Solenoid detector
Author(s)	Panwar, Lata (Bangalore, Indian Inst. Sci.)
Publication	2022 - 301.
Thesis note	PhD : Bangalore, Indian Inst. Sci. : 2022
Thesis supervisor(s)	Komaragiri, Jyothsna Rani
Subject category	Detectors and Experimental Techniques
Accelerator/Facility, Experiment	CERN LHC ; CMS
Abstract	Since the discovery of the Higgs boson in 2012 at CERN, the Standard Model (SM) has become a validated theory to understand the fundamental physics of the universe. All of the SM parameters are now established as a result of this discovery. Despite this success, it does not explain several observed phenomena such as dark matter, hierarchy problems, and neutrino masses. This motivates to explore new theories beyond the SM (BSM), which could explain these phenomena while preserving the achievements of SM. Apart from this, Higgs boson self- coupling is yet to be determined precisely as Higgs boson pair production cross section is too low to measure it with current experimental potential. Nevertheless, its range constrained by current measurements allows the possibility of Higgs self-coupling interaction via BSM physics. The Large Hadron Collider (LHC) at CERN is the largest hadron collider, where hadron-hadron collisions occur at high energies. It plays a vital role in the search for new physics and SM precision measurements, including the measurements of Higgs trilinear self-coupling. For the thesis work, we use LHC data collected by Compact Muon Solenoid (CMS) experiment. It is a general-purpose detector at LHC, collecting hadron-hadron collision data to study a wide range of physics phenomena. In the central part of the CMS experiment, a superconducting solenoid of 6 m internal diameter provides a magnetic field of 3.8 T. A silicon pixel and strip tracker, a lead tungstate crystal electromagnetic calorimeter, and a brass and scintillator hadron calorimeter are found within the solenoid volume, each consisting of a barrel and two endcap parts. Forward calorimeters expand the barrel and endcap pseudorapidity coverage of the detector. Outside the solenoid, muons are detected in the muon gas-ionization chambers embedded in the steel flux-return yoke. During the collision, a two-tiered trigger scheme is used to select events of physics interest. The first level (L1), made up of custom hardware processors, selects events at a rate of about 100 kHz within a time interval of less than four microseconds using data from calorimeters and muon detectors. The second stage, named the high-level trigger (HLT), consists of a farm of processors that run a speed-optimized version of the complete event reconstruction process and reduces the event rate to about 1 kHz before data storage. Using the information from the various elements of the CMS detector, the object reconstruction and identification algorithms recreate and classify each particle in an event. The identified particles are the final objects to carry forward any analysis study. As the last discovered piece of the SM, we use the Higgs boson as a handle to explore BSM predictions. The thesis focuses on the di-Higgs production searches, where at LHC, a pair of Higgs bosons gets produced in the proton-proton collision. We study both resonant and non-resonant di-Higgs production modes. The non-resonant Higgs boson pair production is the only possible process within SM to study the shape of Higgs potential. Besides, the effective field theory (EFT) approach to study low energy signatures of physics existing at a high energy scale is also suitable for such a production mechanism. We explore BSM resonances for resonant di-Higgs production mode, predicted by the warped extra dimension model and the next-to-minimal supersymmetric model that directly couple with the SM-like Higgs boson. Therefore, it is more accessible to perform direct resonant searches, which could improve the overall sensitivity of di-Higgs searches at LHC. In the first part of the research work, we perform di-Higgs searches at the high luminosity LHC (HL-LHC) in the final state of four b quarks at 14 TeV center-of-mass energy. With a vast amount of data up to 3 ab$^{−1}$ of the total integrated luminosity, the HL-LHC would allow investigation of known SM mechanisms with high accuracy and rare new particles. The study is made using simulations of phase-2 CMS detector assuming multiple proton-proton collisions (up to 200) within each bunch-crossing. We start with the analysis for the vector boson fusion (VBF) resonant di-Higgs production in a boosted regime. The resonance is a massive spin-2 bulk KK graviton particle predicted by the warped extra dimension model. Previous s-channel searches made by the ATLAS and CMS collaborations for these resonances observed no deviation from the SM predictions, which indicates that the resonances might not directly couple with SM quarks and gluons. This inspires us to study the massive resonance production in the VBF production mode as quark-quark fusion starts contributing equally at high energies, thus enhancing the VBF production cross section. The analysis is the first one that explores the VBF resonant production mechanism. The Higgs bosons coming from the decay of massive resonance must be sufficiently Lorentz-boosted in order to reconstruct them as a large-area jet. The signal also contains two energetic VBF jets in the forward pseudorapidity regions of the detector. Despite having the largest branching fraction with four final state $b$ quarks among various di-Higgs modes, the channel suffers from high contamination of SM multijet processes. The unique topology of the production and decay would benefit from the upgraded phase-2 CMS detector, which features an extended tracker coverage to identify b quark originated jet and increase signal acceptance. On the other hand, a high granularity calorimeter in the forward pseudorapidity region will help to identify energetic VBF jets in the signal from the background jets. The SM multijet processes are considered as the main background. We optimize event selection for signal topology and the algorithm to identify $b$ quark-originated jets inside the large-area jet from the Higgs boson. The systematic uncertainties are taken following the CMS recommendations for HL-LHC conditions. Expected signal significance for observation of a bulk KK graviton, having a mass between 1.5-3.0 TeV and up to 5% narrow-width, is projected, assuming 1 fb signal cross section and considering di-Higgs invariant mass as an observable. Following a similar boosted analysis strategy, non-resonant di-Higgs production for the SM and effective field theory (EFT) motivated shape benchmarks are also studied in the same final state. A 95% confidence level (CL) upper limit on the product of Higgs boson pair production cross section and branching fraction is presented for the benchmarks. This boosted strategy has not been proven optimal for the SM benchmark. However, the results project significant sensitivity for EFT motivated non-resonant di-Higgs production benchmarks at the HL-LHC. In the second part, we use the 2016, 2017, and 2018 LHC Run-2 period data collected by the CMS detector at 13 TeV center-of-mass energy with 137 fb$^{−1}$ total integrated luminosity and present the study for resonant di-Higgs production via gluon-gluon fusion in the final state of two photons and two bottom quarks ($b$$\bar{b}$$\gamma$$\gamma$) in a resolved regime. The physics is motivated by the warped extra dimension model where spin-0/2 resonance decays into two Higgs bosons and the next-to-minimal supersymmetric model where spin-0 resonance decays into a Higgs boson and another spin-0 particle different from the discovered Higgs boson. This is the first analysis in CMS collaboration, which explores an NMSSM motivated scenario in $b$$\bar{b}$$\gamma$$\gamma$ state. The analysis benefits from the excellent energy resolution of the CMS electromagnetic calorimeter and good trigger efficiency, which improves the invariant mass of the diphoton system resulting from the decay of the Higgs boson. This channel is the most sensitive among the di-Higgs decay modes, especially for low resonance masses. The data-driven diphoton QCD background and simulated single Higgs production processes are used as the main backgrounds. The analysis uses machine learning methods and validates them to reject these background contamination. It increases the analysis sensitivity despite having a low di-Higgs branching fraction channel. For signal extraction, the fit is performed in a two-dimensional mass plane of diphoton and dijet invariant mass observables. The impact of systematic uncertainties on final results is found to be around 1-2%. With the narrow-width approximation, a model-independent 95% CL upper limit on the product of resonant di-Higgs production cross section and branching fraction is set for resonance mass up to 1 TeV. The results are also compared with appropriate BSM predictions to exclude allowed resonance mass ranges.