Measurement of the tt̄γ production cross section with the ATLAS experiment at √ s = 13 TeV using the full Run 2 data set
by Andreas Achim Kirchhoff
Date of Examination:2024-03-14
Date of issue:2024-12-19
Advisor:Prof. Dr. Arnulf Quadt
Referee:Prof. Dr. Arnulf Quadt
Referee:PD Dr. Carmen Diez Pardos
Persistent Address:
http://dx.doi.org/10.53846/goediss-10896
Files in this item
Name:Dissertation.pdf
Size:33.3Mb
Format:PDF
The following license files are associated with this item:
Abstract
English
With Run 1 of the LHC, an unequalled top-quark factory started its operation. While during Run 1 the focus was on tt̄ measurements, the unprecedented size of the ATLAS Run 2 data set of 140 fb −1 allows for precise measurements even for rare processes involving top quarks. One of these is the tt̄γ process, where a pair of top quarks is produced in association with a photon. This process is sensitive to the electromagnetic coupling of top quarks and hence to the charge of the top quark. Besides the general background of tt̄ measurements that come with an additional photon, the largest irreducible background comes from those events where the emission of the photon happens in the decay of the top quarks. This reduces the sensitivity to the top-photon coupling. Hence, it is desirable to differentiate between photons emitted by top quarks (also called the production part) and photons emitted during the decay. Strictly speaking, there is an interference between these two processes but within the narrow-width approximation, which is valid for the tt̄γ process, it vanishes. Consequently, this analysis for the first time performs a cross-section measurement only on the tt̄γ from production part, where initial state radiation from q q̄ initiated events is not decoupled. The inclusive cross-sections are measured in the dilepton and lepton+jets decay channels of the tt̄ pair. The separation of tt̄γ from production and tt̄γ from decay is done with feed-forward neural networks, where a multiclass classification process was implemented for the lepton+jets channel, while in the dilepton channel binary classification was used. The signal strength and consequently the cross- sections are extracted with a profile likelihood fit. The result for the combined phase space of dilepton and lepton+jets channel at particle level is 322 +16 −15 fb = 322 ± 15(syst.) ± 5(stat.) fb and is in agreement with tWith Run 1 of the LHC, an unequalled top-quark factory started its operation. While during Run 1 the focus was on tt̄ measurements, the unprecedented size of the ATLAS Run 2 data set of 140 fb −1 allows for precise measurements even for rare processes involving top quarks. One of these is the tt̄γ process, where a pair of top quarks is produced in association with a photon. This process is sensitive to the electromagnetic coupling of top quarks and hence to the charge of the top quark. Besides the general background of tt̄ measurements that come with an additional photon, the largest irreducible background comes from those events where the emission of the photon happens in the decay of the top quarks. This reduces the sensitivity to the top-photon coupling. Hence, it is desirable to differentiate between photons emitted by top quarks (also called the production part) and photons emitted during the decay. Strictly speaking, there is an interference between these two processes but within the narrow-width approximation, which is valid for the tt̄γ process, it vanishes. Consequently, this analysis for the first time performs a cross-section measurement only on the tt̄γ from production part, where initial state radiation from q q̄ initiated events is not decoupled. The inclusive cross-sections are measured in the dilepton and lepton+jets decay channels of the tt̄ pair. The separation of tt̄γ from production and tt̄γ from decay is done with feed-forward neural networks, where a multiclass classification process was implemented for the lepton+jets channel, while in the dilepton channel binary classification was used. The signal strength and consequently the cross- sections are extracted with a profile likelihood fit. The result for the combined phase space of dilepton and lepton+jets channel at particle level is 322 +16 −15 fb = 322 ± 15(syst.) ± 5(stat.) fb and is in agreement with the standard model prediction. Furthermore, a summary of differential measurements, an EFT interpretation and a combination with the ATLAS tt̄Z measurement are given.he standard model prediction. Furthermore, a summary of differential measurements, an EFT interpretation and a combination with the ATLAS tt̄Z measurement are given.
Keywords: particle physics; high-energy physics; ATLAS experiment; Standard Model; top quark; top-quark properties; top-photon coupling; cross-section measurement; fiducial cross-section; machine learning; top-quark physics
