Subtracting Stellar Signals and Improving Instrumentation On A Road Towards Other Earths

摘要

Improvements in radial velocity precision and detection capabilities will one day lead to the detection of earth-mass planets orbiting in the habitable zones of nearby stars. Detecting low mass planets at large separations will help fill in our understanding of how planetary systems form and subsequently evolve. Additionally, they will provide a target list for followup with future facilities such as the James Webb Space Telescope that will search for atmospheric biosignatures and help determine how common life is in the universe. This thesis presents observations, methods, and analyses created to improve the instrumental precision, efficiency, and detection capabilities of the radial velocity (RV) method of exoplanet detection.;Chapters 2 and 3 present 7 planetary systems discovered using the HIRES Spectrograph on the Keck I telescope. One of these systems presents a challenge for theorists, as it cannot be explained with current planet formation and migration theories. We also detect a correlation between RV measurements and S-value measurements, which are an indicator of stellar activity. By linearly modeling and removing this correlation, the residual RMS of the velocities improved by 21%.;In Chapters 4 and 5 we describe the commissioning and software stack behind CHIRON, a new instrument we commissioned on a 1.5 m telescope at the CTIO Observatory. CHIRON provides a queue-based observing system for the general astronomical community, and dealing manually with the large amount of data collected nightly quickly became an overwhelming task. Primarily to reduce time and error, most of the observing pipeline for CHIRON was automated, which is described in these two chapters.;Lastly, we present observations from a multi-facility campaign to simultaneously observe the young and moderately active K dwarf e Eridani. We use this data to explore three methods of modeling stellar activity: spot modeling, the FF' method, and a new approach that makes use of Ha measurements input into the FF' method, which we refer to as the HH' method. Spot modeling yielded the best result of the three methods tested in terms of residual RMS after subtracting the model from the data, and reduced the RMS from 7.4 m s -1 to 2.7 m s-1. However, the HH' method benefits from not needing precise space-based photometry as input. Additionally, the HH' method only makes use of one of the thousands of lines in the visible spectrum; incorporating more spectral lines into the model would most likely lead to significant improvements.