Master's Research

Research Overview

Particle identification (PID) in a nuclear physics experiment is, oftentimes, essential to the physics analyses that utilize the data collected by such experiments. This is also true for my PhD research. Therefore, part of the motivation for my master's research is to set the stage for PID needed for my PhD research. STAR primarily relies on dE/dx energy loss information from the Time Projection Chamber (TPC) for PID. Measurements from the time-of-flight (TOF) system at STAR can be used to provide PID information when dE/dx PID information has a difficult time distinguishing between particle types. For my purpose, I improved the existing “start-less” TOF algorithm, which infers the collision time without reference to a separate start time detector. Prior to my work, start-less TOF had only been used when analyzing low-energy Au+Au collisions at STAR. Now, it can be used for p+p, p+Au, and fixed target analyses.

There were a few issues with the default start-less TOF algorithm. The default pion selection cut for start-less TOF is too generous, so I developed a more restrictive set of cuts for pion selection that yields a higher purity pion sample. When looking at the default outlier rejection algorithm for start-less TOF, I found that it is a single-pass rejection loop, where the hits that are rejected depend on the order in which the hits appear in the list. Another problem was that the algorithm accepts a very wide range of nominal collision times. In the improved outlier rejection algorithm, I address both issues. We also introduced a momentum-dependent dE/dx time correction to correct the particle time-of-flight calculations based on energy loss due to traveling through the material inside of STAR. Other modifications include a new set of quantities for particle identification and new tools to identify events that are contaminated by in-bunch pile up.

When comparing our optimized start-less TOF to default start-less TOF for p+Au in the figure below, we see that the integrated yield is increased by ~2%, the RMS decreases by ~19%, and the single-hit time resolution is σBTOF ~ 75 ps compared to the default resolution of ~87 ps. Similar results are observed in p+p. Also, the dE/dx correction shifts the centroid of the plot in the figure below closer to zero, meaning the measured flight time matches what is expected for a pion of that momentum. All the improvements I made have been added to the STAR software library, and have been utilized by other researchers in the collaboration.