Christopher Hearty

Even though the Standard Model of particle physics is a very successful model, we know that it is incomplete. For example, it does not explain why the world we see around us is made of matter rather than anti-matter and it does not incorporate the force of gravity. One extension of the Standard Model is the introduction of Axion Like Particles, ALPs. ALPs appear in string theories and supersymmetry and they might explain some astrophysical anomalies. ALPs can be produced in electron-positron colliders and detected in the specialized detectors built around their interaction point, like the PEP-II collider and the BaBar detector at the Stanford Linear Accelerator Center. This work presents an un-blinded search for an ALP that couples exclusively to photons in 5% of the BaBar data. We search for an excess in the invariant mass distribution of ALP candidates over a smooth background. The results are consistent with the data being composed only of Standard Model background. 90% credible interval upper limits are set on the ALP production cross section and coupling constant. These limits exclude previously unexplored regions of the phase space in the mass range 0.29 GeV/c² to 5 GeV/c². In searches involving photons, it is important to be able to efficiently detect them while rejecting other types of particles. Many high energy particle detectors detect photons in electromagnetic calorimeters that are made up of many cells. A photon interacting with the calorimeter typically leaves a different energy distribution in the cells than some other particle types, hadrons, for example. Discriminating variables for photons, based on Zernike moments, are developed in order to improve the photon identification at Belle II. One of the new variables is found to be the best at identifying photons among all other such variables used at Belle II for photons with energies in the energy range most relevant to e+e− → BB events.

View record

Drift chambers are a type of gaseous ionization detector used in high-energy physics experiments. They can identify charged particles and measure their momentum. When a high-energy charged particle crosses the drift chamber, it ionizes the gas. The liberated electrons drift towards positive-high-voltage wires where an ionization avalanche amplifies the signal. Traditional drift chambers use only the arrival time of the cluster of charge from the closest ionization for tracking, and use only the integral of the whole signal for particle identification. We constructed prototype drift chambers with the ability to resolve the charge cluster signals from individual ionization events. Different algorithms were studied and optimized to best detect the clusters. The improvementsto particle identification were studied using a single-cell prototype detector, while the improvements to particle tracking were studied using a multiple-layer prototype. The prototypes were built in the context of initial work for the now-cancelled SuperB project, but the results apply to any drift chambers used in flavour-factory experiments. The results show that the choice of algorithm is not as critical as properly optimizing the algorithm parameters for the dataset. We find that a smoothing time of a few nanoseconds is optimal. This corresponds to bandwidth of a few hundred megahertz, indicating that gigahertz-bandwidth electronics are not required to make use of this technique. Particle identification performance is quantified by the fraction of real pions correctly identified as pions with at most 10% of real pions mis-identified as muons. In our single-cell prototype, the performance increases from 50% to 60% of pions correctly identified when cluster counting is combined with a traditional truncated-mean charge measurement, compared to the charge measurement alone. Tracking performance is quantified by the single-cell resolution: the uncertainty in measuring the distance of charged particle tracks from a given sense wire. In our multiple-layer prototype, the single-cell tracking resolution using traditional methods is measured to be ~150μm. With cluster counting implemented, the resolution is unchanged, indicating that the additional cluster information is not useful.

View record

In the Standard Model of particle physics, the Higgs boson is an elementary particle required to explain the origin of mass. One extension, the next-to-minimal supersymmetric Standard Model, predicts the existence of a GeV/c² scale Higgs boson, denoted A^0. This can be produced at B meson factories such as the BABAR experiment at the Stanford Linear Accelerator Center. We study e⁺e- → Υ(2S)→ π⁺π-Y(1S), Y(1S)→ γA^0, A^0 → gg, ss-bar, cc-bar decays using data collected by the BABAR detector at a centre-of-mass energy of 10.02 GeV/c².The search for A^0 → gg or ss-bar is performed by fully reconstructing the entire decay chain. We search for an excess of events relative to expected backgrounds in the reconstructed A^0 mass spectra. No significant excess is found. We set upper limits on product branching fraction B(Y(1S) → γA^0)xB(A^0 → gg) from 1x10-⁶ to 2x10-² for A^0 masses from 0.50 to 9.00 GeV/c², and B(Y(1S) → γA^0)x B(A^0 → ss-bar) from 5x10-⁶ to 1x10-³ for A^0 masses from 1.50 to 9.00 GeV/c².The search for A^0 → cc-bar is done by reconstructing the dipion transition, the radiative photon, and a D meson as a tag for cc-bar decays. We infer the mass of an A^0 candidate from the recoil of the dipion transition and the radiative photon. No significant signal is seen in the data and we set upper limits on product branching fraction B(Y(1S) → γA^0)xB(A^0 → cc-bar) from 7x10-⁵ to 2x10-³ for A^0 masses from 4.00 to 8.95 GeV/c² and 9.10 to 9.25 GeV/c².These are the first measurements to date and set constraints on the next-to-minimal supersymmetric Standard Model.Proposals to improve drift chamber aging studies and particle identification are included as part of the thesis. We propose to improve aging studies with prototypes that better resemble a full-scale drift chamber. Using TRIUMF beam data, we establish that at a particle identification efficiency of 95%, misidentification rates may be reduced from 2.2% to 1.9% by including likelihood tests and Kolmogorov-Smirnov tests and to 0.5% by including cluster counting.

View record

No abstract available.