Searching for Exoplanets with Radial Velocities from the MINERVA Telescope Array

Abstract

MINERVA uses four 0.7-m telescopes located atop Mt Hopkins in Arizona to observe some of the nearest and brightest stars in the sky. MINERVA employs the use of high-resolution spectroscopy to see if there is any evidence for planetary systems. From the spectroscopy, we have previously extracted the Radial Velocity (RV), or Doppler shift of the star as a function of the times of observations.  The observations span several years and are unevenly sampled in time, making it impractical to use traditional analysis techniques such as the Fourier Transform.  A periodogram is analogous to a Fourier Transform but applicable to unevenly sampled data.  We analyze the RVs for one of these stars. The goal is to take the unsearched data and look for interesting peaks in the periodogram.  Periodogram “power” peaks  are indicative of the presence of a periodic signal that could be due to the stellar reflex motion of an orbiting exoplanet. After successful installation, we first used Julia to run the RVModelingToolKit on a test set of RVs from the already known system KELT-24, which possesses a massive hot Jupiter planet KELT-24 b, and we recover a periodogram peak at the known orbital period of KELT-24 b. We next modified the script to run on the MINERVA RV data sets, along with adding data headers for code compatibility. We plotted and identified outliers in the data RVs that are of lower measurement precision.  These outliers are next trimmed and removed. For any peaks in the periodogram of the trimmed RV data set, we can then assess whether or not these peaks are associated with an exoplanet with an orbital period equal to the period of the peak.  We construct a Keplerian model and compute an MCMC to evaluate the statistical likelihood of whether or not a planet is statistically favored to exist given our planetary model, data, and assumptions.