Detection and reconstruction of short-lived particles produced by neutrino interactions in emulsion

In this dissertation, several different topics related to the chorus experiment are pre- sented. The chorus experiment has been used to study neutrino oscillations using the neutrino beam at cern. The neutrino oscillation hypothesis provided an explanation for the lower than expected fluxes of solar and atmospheric neutrinos. There are three neutrino species in nature corresponding to different weak eigenstates, namely, the elec- tron neutrino (νe ), the muon neutrino (νμ ), and the tau neutrino (ντ ). The lower fluxes could be interpreted as spontaneous oscillations between electron and muon neutrinos and between muon and tau neutrinos. The chorus experiment was designed to detect oscillation of muon neutrinos into tau neutrinos with small mixing probability down to 2 · 10−4 and a mass difference square between νμ and ντ larger than 0.5 eV2 . In the last decade, several disappearance experiments have confirmed the neutrino oscillation hypothesis and showed that oscillations occur between mass eigenstates with mass differences of 8 · 10−5 eV2 and 2.5 · 10−3 eV2 . The observed oscillations correspond to almost maximal mixing probability. Current and future efforts focus on detecting the remaining sub-dominant mixing. In the chorus experiment, one looks for the typical decay signature of a τ particle in a large stack of nuclear emulsion plates. Emulsion can record with very high resolution the tracks and decays of short-lived particles, like τ leptons and charmed mesons. A hybrid setup is used consisting of electronic detectors downstream of a 770 kg emulsion target. The electronic detectors consist of an accurate tracking station followed by a hadron spectrometer, a calorimeter, and a muon spectrometer. These detectors are used to measure particle momenta and to select one track per event for scanning the emulsion. Hybrid detectors combined with automated microscopes have greatly increased the amount of emulsion that can be scanned and the information that can be extracted. The fken laboratory in Nagoya developed designated hardware that can find a track with known angle in a 100 μm slice of emulsion. Further upgraded new hardware can find all tracks with slopes within a given angular range. The hardware is used to locate the neutrino vertex by following an electronically measured track from one plate to the next. The new hardware is then used to reconstruct all tracks and vertices of the event. This is done by combining track segments found in the 100 μm upstream layers of the eight plates around the interaction point. The suppression of background in the chorus experiment relies on the measurement of hadron momenta in the hadron spectrometer. A honeycomb drift-tube tracker was added to the detector to improve this measurement. This made it possible to reconstruct independently 3-d track segments in the hadron spectrometer. This improved the uncer- tainty on hadron momenta from ∆p/p = 0.352 + 0.38p/GeV to 0.352 + 0.25p/GeV, where the constant term is due to multiple scattering. The read-out of this detector is based on a continuous sampling of the over–threshold signal of all wires. Because all transitions of a wire’s signal are recorded, both leading edge and pulse length can be recovered. The bit-stream of samples is stored in four phases in separate memory banks. Failure of some memory banks does not destroy the hit information but only lowers its timing accuracy. summary A new automatic emulsion scanning system was developed at cern. For this, new microscope optics with a large field of view were deployed and emulsion scanning was implemented in software. Using software for track finding in emulsion images requires an algorithm capable of handling of the order of 105 3-d hits in a few seconds. An algorithm was developed that searches a network of connected hits. The connection graph links hits in close proximity with a predefined restriction. In this case, the link should possibly be part of a track. This graph is created by sorting the hits in k -dimensional space using a specially designed ordered container with logarithmic search time. The track-finding algorithm considers any connection between two hits as a possible track and builds longer segments using a limited depth-first search of the connection graph. Although originally written for track finding in emulsion, the track-finding algorithm can work in any dimension and is decoupled from the actual hit and track model. The track model and the acceptance of hits in a track are handled by a specific implementation of an abstract acceptance class. The track-finding implementation has been used to find bent tracks in the harp time-projection chamber. In the chorus emulsion application, it has enabled improvements in the vertex location efficiency. It can also be used to reduce human confirmation of secondary vertices by applying it to data taken over the full emulsion thickness. An emulsion-specific algorithm was developed to look for tracks with a known direction. The algorithm is based on moving hits along the track direction and summing the number of hits in its angular acceptance region. Searching for hits in these acceptance regions can be done in logarithmic time using the aforementioned containers. The track-finding algorithm is then used to verify the existence of a track in the small volumes defined by a track-trigger sum over threshold. The k -d space ordering containers have been used throughout the work described in this dissertation, for example in several alignment procedures. The emulsion volume, scanned around the neutrino vertex, defines a decay space for short-lived particles. The maximum decay length is about 4.5 mm along the beam di- rection. In the chorus experiment, 1048 charged-current νμ interactions, identified by a μ− from the primary vertex, with a secondary decay vertex from a D0 meson have been reconstructed. These events were used to study D∗+ resonance production. The decay D∗+ → D0 π + can be identified using the π + momentum measured in the hadron spectrometer and the direction of both the D0 and π + measured in the emulsion. From the observed 22.1 ± 5.5 events, the cross-section ratio for hadronization of a charm quark into D∗+ and D0 is derived: σ (D∗+ ) /σ D0 = 0.38 ± 0.09(stat) ± 0.05(syst). The production cross-section of D∗+ with respect to the total charged-current cross-section is σ (D∗+ ) /σ (CC) = [1.02 ± 0.25(stat) ± 0.15(syst)] %