Entwicklung von Objekt orientierten C++ Algorithmen zur b-Quark Identifikation

In the past thirty years particle physics has developed rapidly resulting in the formulation of the Standard Model, which seems to provide, at least in principle, a microscopic description for all known physical phenomena except gravity. The Standard Model is not complete, e.g. it lacks any explanation for the pattern of particle masses. The Higgs mechanism provides a solution to the problem of how particles acquire their masses. It implies the existence of at least one new particle, the Higgs boson H0, which has not yet been observed. The Large Hadron Collider (LHC) at the European Organization for Nuclear Research (CERN) will be switched on in winter 2007. If the Higgs boson exists, the LHC will be able to detect it. Depending on the mass of the Higgs boson, physicists have a clear idea regarding its experimental signature. For quite low masses 50 < mH0 < 130 [GeV] the Higgs will predominantly decay into two b-quarks. The present study describes the investigation of the identification capability of b-quark signatures in the CMS experiment at the LHC. In order to test the b-quark identification methods developed, the data of the ALEPH experiment recorded at the LEP e+e- storage ring from 1992 till 1995 were used. The strategies developed for events from e+e- collisions recorded by the ALEPH experiment were applied to simulated pp events in the CMS detector. The new combined b-tag developed for ALEPH developed uses a combination of observables sensitive to the differences of b-hadron decays from other light quark hadron decays. The b-tag is based on the the ALEPH reconstruction algorithms for tracks and jets optimized for b-hadron identification. It is provided by the QIPBTAG tool of the ALEPH reconstruction tools library ALPHA. QIPBTAG also provides the calibrated resolution function obtained from the significance of the signed 3-dimensional impact parameter and used this to compute the track probabilities and the jet lifetime probability. The developed b-tag for ALEPH was put in perspective by comparison with the latest Rb measurements at the Z0 resonance: e.g. The newly implemented b-tag is able to operate in the same b-efficiency and b-purity region as the latest DELPHI measurement, which is the actual single best measurement of Rb. For the same b-purity used in the measurement of Rb with the ALEPH lifetime mass tag, the b-efficiency increases by roughly about 10% for the newly developed b-tag. To be able to use a similar combined b-tag for CMS the prerequisites were created. The implementation of a 3-dimensional jet lifetime probability b-tag for CMS events required the optimization of the CMS track, vertex and jet reconstruction. The jet reconstruction was performed by the inclusive kT jet algorithm using a newly developed jet energy flow input (TrackTowers). This jet reconstruction setup is able to provide reasonably determined final state jets in the dense event topologies of the LHC. Because of the occurrence of additional primary vertices in LHC pp collisions, e.g. due to pileup interactions, strategies were developed to reconstruct the primary vertices with high accuracy and to reliably identify the signal vertex of the main interaction. The selection of tracks from the primary b-hadron decay vertex was done by implementing constraints to suppress other tracks, e.g. from other jets, other primary or tertiary vertices. Having optimized the reconstruction and selection of tracks, vertices and jets as described above, the signed 3-dimensional impact parameter was used to compute the resolution function, which is needed to calculate the probability of an individual track to stem from the signal vertex, and the jet lifetime probability, which combines the individual track probabilities to a single observable. A good performance of the jet lifetime probability b-tag was obtained for typical LHC t anti-t events mixed with low-luminosity pileup (L = 2 x 1033 cm-2s-1). For the implementation of the combined b-tag in CMS, a search for a secondary vertex was performed inside the reconstructed jets. As for ALEPH, the association of tracks to the secondary vertex was done by comparing the track distance from the secondary to that from the signal vertex. Because of the huge amount of gluon jets in LHC events the combination of the used observables was revised and a separate weight for the gluon-jets was utilized in the CMS case. The b-hadron jet identification was significantly improved by the combined b-tag with respect to the jet lifetime probability b-tag alone. Finally the discovery potential of CMS for the WH, H -> b anti-b channel in low-luminosity pileup conditions was investigated by employing the improved and newly developed CMS reconstruction algorithms. The discovery of the Standard Model Higgs boson in this associated channel is almost impossible in the first years of LHC even though the W mass constraint is utilized, because of the large cross section of the t anti-t production and the very similar decay topologies of Wbb events in comparison to the WH production. The investigation and development of the presented b-hadron jet identification for ALEPH and CMS was done by means of a newly implemented, purpose-built, stand-alone package based on the ROOT object orientated data analysis framework. To decouple the package from the ALEPH reconstruction tools library (ALPHA) and the CMS reconstruction framework (ORCA), the reconstructed events were preselected and stored in a common data format. The full simulation, full reconstruction and pre-selection of the CMS data was done by means of the Worldwide LHC Computing Grid. The data from the LHC experiments (roughly 15 Peta bytes per year) will be distributed around the globe so that thousands of scientists around the world will be able to access and analyze it. For the present study, CMS Monte Carlo samples of 2M events were reconstructed dozens of times. The CPU time needed per event is several minutes. Thus, in addition to developing new physics tools and algorithms, the methods used reflect the global way in which analyses of CMS data will be performed.

Grantor dc:publisher

Publikationsserver der RWTH Aachen University

Year dc:date

2007

Author dc:creator

• Heister, Arno

Contributors dc:contributor

• Schael, Stefan

• info:eu-repo/classification/ddc/530
• Higgs-Modell
• Higgs-Teilchen
• Higgs-Mechanismus
• ALEPH <Teilchendetektor>
• CMS-Detektor
• LHC
• b-Quark
• LEP
• Physik
• b-Tag

Statement dc:rights

• info:eu-repo/semantics/openAccess

Language dc:language

ger

Identifier

https://publications.rwth-aachen.de/search?p=id:%22RWTH-CONV-124106%22

OAI identifier oai:identifier

oai:publications.rwth-aachen.de:62542