Searches for vector-like quarks with 13 TeV at the ATLAS experiment and development of a boosted-object tagger using a deep neural network
| dc.contributor.advisor | Kröninger, Kevin | |
| dc.contributor.author | Freundlich, Elena M. | |
| dc.contributor.referee | Rhode, Wolfgang | |
| dc.date.accepted | 2020-09-04 | |
| dc.date.accessioned | 2020-11-04T07:53:31Z | |
| dc.date.available | 2020-11-04T07:53:31Z | |
| dc.date.issued | 2020 | |
| dc.description.abstract | In this thesis, two searches for the pair production of vector-like top (VLT) and vector-like bottom (VLB) quarks are presented using 36.1 fb⁻¹ and 139 fb⁻¹ of data. The data was taken at √s = 13 TeV in proton-proton collisions with the ATLAS detector, which is one of the major experiments at the Large Hadron Collider at CERN. The decay signature of this search for TT and BB production is characterized by the presence of a Z-boson, which is reconstructed from two opposite-charge same-flavor leptons (e or μ) and is produced alongside a third-generation quark. In addition, the analyses target final-state objects with high transverse momentum, pT. Within the scope of this thesis, the Multi-Class Boosted-Object Tagger (MCBOT) is developed and characterized. It uses a deep neural network, which identifies reclustered large-radius jets originating from V-bosons, H-bosons, top-quarks and background unambiguously. For the search with 139 fb⁻¹, the multiplicity of hadronic high-pT objects is exploited by defining multiple regions according to the number of MCBOT- and b-tags. No deviation from the Standard Model expectation is found in the observed data for 36.1 fb⁻¹ and 139 fb⁻¹, and exclusion limits on the pair production cross-section and the mass of VLT and VLB quarks are set depending on various branching-ratio assumptions. With 36.1 fb⁻¹, VLT quark masses below 1102 GeV are excluded with respect to the (T,B) and (X,T) doublet models. For VLB quarks, lower mass limits of 1063 GeV are set for the (B,Y) doublet. The results of this search significantly contribute to the sensitivity of two combination searches for TT and BB production performed at the ATLAS experiment with 36.1 fb⁻¹, which constrain the phase-space for vector-like quark existence further. For the full Run-2 search with 139 fb⁻¹ of data, lower limits on the VLT quark mass of 1350 GeV are observed and VLB quark masses below 1316 GeV are excluded assuming the mentioned doublet models. The observed limits of this search for TT and BB production represent the highest exclusion limits to date assuming the 100% branching ratio into Zt/b final states and – in case of VLB quarks – for the (B,Y) doublet model. | en |
| dc.identifier.uri | http://hdl.handle.net/2003/39803 | |
| dc.identifier.uri | http://dx.doi.org/10.17877/DE290R-21694 | |
| dc.language.iso | en | de |
| dc.subject | ATLAS | en |
| dc.subject | Vector-like quarks | en |
| dc.subject | Machine learning | en |
| dc.subject | Deep neural networks | en |
| dc.subject | MCBOT | en |
| dc.subject | LHC | en |
| dc.subject | New physics | en |
| dc.subject | Particle physics | en |
| dc.subject.ddc | 530 | |
| dc.subject.rswk | ATLAS <Teilchendetektor> | de |
| dc.subject.rswk | Quark <Physik> | de |
| dc.subject.rswk | Deep Learning | de |
| dc.subject.rswk | Neuronales Netz | de |
| dc.subject.rswk | Maschinelles Lernen | de |
| dc.subject.rswk | LHC | de |
| dc.subject.rswk | Elementarteilchenphysik | de |
| dc.title | Searches for vector-like quarks with 13 TeV at the ATLAS experiment and development of a boosted-object tagger using a deep neural network | en |
| dc.type | Text | de |
| dc.type.publicationtype | doctoralThesis | de |
| dcterms.accessRights | open access | |
| eldorado.secondarypublication | false | de |