Searching for Extragalactic Exoplanets: A Survey of the Sagittarius Dwarf Galaxy Stream with TESS

Abstract

To date no exoplanets have been detected outside the Milky Way, and their extragalactic occurrence rates are poorly constrained. Using available data from TESS we perform the first transit survey of the Sagittarius dwarf galaxy stream using 15,176 main sequence stars identified as likely members. We calculate an upper limit of $<$1.01% for hot Jupiters with radii of 1-2 R$_{Jup}$ and periods of 0.6-10 days after detecting zero planets. We compare our calculated occurrence rate upper limits to the upper limits found in the Milky Way globular clusters M4 and 47 Tuc. Our 1-$σ$ occurrence rate upper limit of $<$0.37% for the Sagittarius dwarf galaxy stream, for planets with radii of 1.5-2 R$_{Jup}$ and periods $<$10 days, is lower than the $<$0.57% upper limit measured in 47 Tuc. Similarly, our 2 sigma upper limit of $<$0.78% for planets with radii of 1.4-2 $_{Jup}$ and periods $<$8 days is below the $<$0.81% upper limit measured in M4. We predict that a future analysis of TESS data with a high detection efficiency for hot Jupiter transit depths would require $η_{extragalactic}$ $\geq$ 11,467 target stars to detect a planet of extragalactic origin. Therefore, we predict that a future investigation of TESS data which includes additional extragalactic stellar streams will be able to either detect the first extragalactic origin planet or provide evidence that older, lower metallicity extragalactic environments may experience a lower hot Jupiter occurrence rate than is observed for the Milky Way.

Figures (8)

• Figure 1: Comparison between the selected T$_{mag}$=16 normalized median absolute deviation value used as an anchor point for our Poisson-like noise cutoff, and the number of rejected stars at that noise cutoff. At a normalized median absolute deviation anchor point of 0.0065 the function becomes linear out to brighter median absolute deviation cutoffs (shown in black). Due to significant contamination from nearby bright neighboring stars the normalized median absolute deviation values of some target stars fall below the expected noise limit. The transition at 0.0065 represents the point where our noise cutoff removes these poorly fit light curves.
• Figure 2: The normalized median absolute deviation of each light curve in our survey as a function of T$_{mag}$ where the 9,768 light curves reduced with TGLC are shown blue, and the 5,408 light curves reduced with eleanor are shown in green. For the faint target stars stars observed in the most crowded fields of our survey, significant contamination from nearby bright neighboring stars can lead to light curves with median absolute deviation values that fall below the predicted noise limit (Feinstein2019PASP..131i4502F. To remove these highly contaminated light curves we plotted the number of stars rejected at a series of Poisson-like noise cutoffs and selected the cutoff where the function began to deviate from the linear function seen at brighter median absolute deviation cutoffs (see Figure \ref{['cutoff']}). We show this selected noise cutoff as an orange curve in the scatter plot.
• Figure 3: (a) The stellar radius distribution of the stars from ramos22 with our cuts of R$_*$$<$2 R$_{\odot}$ and Sagittarius membership probability $>$50$\%$ across the full magnitude range of G $<$ 19.75. The median stellar radius of this distribution is 1.04 R$_\odot$. (b) The stellar radius distribution for the 15,176 stars in our survey with our cuts of R$_*$$<$2 R$_{\odot}$, Sagittarius membership probability $>$50$\%$, and T$_{mag}$$<$17. The median stellar radius of this distribution is 1.38 R$_{\odot}$.
• Figure 4: HR diagram depicting our sample of 15,176 probable Sagittarius dwarf galaxy stream members in blue against the 7,993 planet host stars from the Kepler, K2, and TESS transit surveys Berger2023arXiv230111338B in orange. The sharp cutoffs seen at higher luminosities are the result of magnitude limit cuts in the initial ramos22 stellar sample as well as our own cuts in stellar radius and T$_{mag}$. These cuts were used to reduce the number of evolved stars in our sample.
• Figure 5: A 15x15 Full Frame Image cutout centered on the target star TIC 92223525 (green star). The red dots represent the neighboring stars in the Gaia DR3 catalog with T$_{mag}$$<$17, and the blue star represent the neighboring star TIC 92223526 we identified as the source of the observed candidate transit signal. With a separation of just $\sim$2 TESS-pixels, we estimate that the light curve of TIC 92223525 was contaminated by 11.5% of the observed flux from the neighboring star TIC 92223526 which we believe to be an eclipsing binary.