Planetary Architectures of Kepler Compact Multis with Binary Star Companions

Abstract

Planets in binary-star systems exhibit demographic differences compared to planets in single-star systems. In particular, planets with binary-star hosts have a lower overall occurrence rate compared to their single-star counterparts, as well as a suppressed relative occurrence rate for sub-Neptunes ($R_p=2{-}4R_{\oplus}$) compared to super-Earths ($R_p=1.0{-}1.5R_{\oplus}$). These differences are most pronounced in close separation binaries ($ρ< 100$ au) which has been interpreted as a result of binary stars disrupting the protoplanetary disks of their stellar companions. The architectures of planetary systems -- i.e. the arrangements of planet sizes and orbits -- provide additional information about system formation and evolution. Architectures of single-star planetary systems are well studied, but architectures of binary-star planetary systems have not been investigated in detail. In this work, we analyzed a large sample of Kepler planetary systems (162 planets in 118 binary-star systems; 880 planets in 544 single-star systems) to compare their architectures as a function of stellar multiplicity. We found that planets with binary-star hosts follow a similar ``peas-in-a-pod'' tendency toward uniformity in planet radii and log-uniformity in period spacing as planets with single-star hosts. However, we also detected modest ($2.5-3σ$) differences in period spacing and planet multiplicity, with binary-star systems having higher typical gap complexities (indicating more uneven spacing) and a higher prevalence of single planets. We interpret these results as evidence that binary stars primarily influence the planetary architectures of their stellar companions by shaping the protoplanetary disk at formation, with subsequent dynamical processing more gently altering the system architectures over secular timescales.

Figures (15)

• Figure 1: Period-radius plot for planets in single-star (gray points) and binary-star (colored points) systems considered in this work. Colors correspond to log$_{10}(\rho)$, i.e. the projected physical binary separation in au. We restricted the sample to planets with $1 < P < 100$ d and $0.5 < R_{p} < 4 R_{\oplus}\xspace$. The planets in binary-star systems have a similar distribution in period-radius space as the planets in single-star systems, although the relative occurrence rate of sub-Neptunes ($2 < R_p < 4 R_{\oplus}\xspace$) is suppressed in small-separation binary-star systems Sullivan2024.
• Figure 2: 2-planet systems with binary-star hosts. Systems are ordered by binary separation, with close separations at the bottom and wide separations at the top. All $R_{p}$ assume the planets orbit the primary star. Earth, Neptune, and Jupiter radii (1, 4, and 10 $R_{\oplus}$) are shown in the top row for comparison.
• Figure 3: $3+$ planet systems with binary-star hosts. Systems are ordered by binary separations, with close separations at the bottom and wide separations at the top. All $R_{p}$ assume the planets orbit the primary star. Earth, Neptune, and Jupiter radii (1, 4, and 10 $R_{\oplus}$) are shown in the top row for comparison.
• Figure 4: Fraction of planetary systems with 1, 2, 3, or 4+ planets detected for single stars (black circles), all binary stars (pink triangles), close binary stars (blue triangles) and wide binary stars (yellow triangles). The fraction of single-planet systems is larger for all binary-star planet host cases, and is statistically significant (3.0$\sigma$) based on a $\chi^{2}$ contingency test for all stellar binaries versus stellar singles.
• Figure 5: Left: Cumulative distributions of the outer/inner radius ratios $R_p'/R_p$ for adjacent planet pairs in binary-star systems (pink) and single-star systems (black). Right: Cumulative distributions of planet radius Gini indices $G_R$. The faint lines show the scatter of the distributions, calculated by drawing new $R_p\xspace/R_{\star}\xspace$ randomly from the normally distributed $R_p\xspace/R_{\star}\xspace$ errors and recalculating $R_p'/R_p$ and $G_R$; for visual clarity, we display these uncertainty bootstraps only for the binary-star systems and not the single-star systems. The binary-star systems and single-star systems are statistically indistinguishable within 1$\sigma$ for both $R_p'/R_p$ and $G_R$, suggesting that planets in binary-star systems have the same degree of size self-similarity as planets in single-star systems.