Do Jupiters Hurt the Formation of Earths?

Title: Cold Jupiters and improved masses in 38 Kepler and K2 small-planet systems from 3661 high-precision HARPS-N radial velocities. 

Authors: A. S. Bonomo, X. Dumusque, A. Massa, A. Mortier, R. Bongiolatti, L. Malavolta, A. Sozzetti, L. A. Buchhave, M. Damasso, R. D. Haywood, A. Morbidelli, D. W. Latham, E. Molinari, F. Pepe, E. Poretti, S. Udry, L. Affer, W. Boschin, D. Charbonneau, R. Cosentino, M. Cretignier, A. Ghedina, E. Lega, M. López-Morales, M. Margini, A. F. Martínez Fiorenzano, M. Mayor, G. Micela, M. Pedani, M. Pinamonti, K. Rice, D. Sasselov, R. Tronsgaard, and A. Vanderburg

First Author’s Institution: Osservatorio Astrofisico di Torino

Status: available on the arXiv [open access]

When looking at exoplanet systems, a conundrum appears: if so many exoplanet systems have sub-Neptune planets (< 4 Rearth) in short (< 100 day) orbits, why is our Solar System missing this kind of planet? Astronomers believe that Jupiter-like planets may be the culprit. These planet’s large masses may have either acted as a “dynamical barrier” to the migration of sub-Neptunes from their cold formation zones at the edges of the system to an orbit closer to their star. Or, if these planets form in place, Jupiter-sized planets may have prevented material needed for planet formation from ever even reaching the inner system.

We can begin to test whether Jupiters really hurt the formation of these smaller, interior sub-Neptunes by studying exoplanetary systems for clues. By looking at how often small planets occur, how often Jupiters occur, and how often they occur together, we can get a sense of whether or not the two populations are related and form together. On the other hand, if the theories about Jupiter’s hindrance of small planet growth are true, we can also expect to see an anti-correlation between the two populations; that is, if there is a Jupiter then there is no inner sub-Neptune and vice-versa. This is exactly what today’s authors want to find out. They report the results of a decade+ long survey set to answer the question: how often to Jupiter-like planets occur in systems with known, small, inner planets? 

The survey was designed using the HARPS-N spectrograph on the Telescopio Nazionale Galileo located on La Palma in the Canary Islands. The team carefully selected 38 exoplanet systems that are known to have small planets on short orbits from the Kepler and K2 missions. Then, they monitored all 38 stars for 10 years, collecting a few Radial Velocities (RVs) each year. Normally, when trying to measure known planets’ masses, surveys will want to maximize the number of RVs they can collect within a relatively short amount of time, say a year or two. But this survey was not looking to measure masses. Rather, it was looking for the tell-tale sign of a Jupiter-like planet. When a planet has a very long orbital period, like Jupiter’s 12 year orbital period, and you only sample the RVs over a shorter timespan, you won’t see a full orbit. Instead, you will see a trend in the RV data set. Take a sine wave, like in the left side of Figure 1, which represents the periodic change in the velocity of a star due to a planet’s gravitational influence. In one full orbital period of the planet, you will see the rise and fall of the velocity of the star. But if you only sample part of the velocity wave, like in the right side of Figure 1, you’ll see only a line, or a trend in the data. This is exactly what the team was looking for in these stars: trends that represent partial orbits of very long period planets. 

Figure 1: On the left is a representation of RV monitoring of a star over the course of a full orbit from a giant planet. Over a full orbit, the velocity of the star rises and falls as the planet orbits the star. If we only observe for a relatively short time, represented by the time within the red box, then the velocity curve looks identical to a straight line trend. Therefore, trends in RV data sets are indicative of planets with orbital periods much longer than the baseline of observations. Image Credit: Jack Lubin.

Out of the 38 stars the team followed over 10 years, 31 did not show any signs of Jupiter-sized planets. Of the remaining 7 systems, 3 show evidence for a full orbit of a giant planet and 5 systems show evidence for a partial orbit trend (one system has both a full orbit and a trend). However, when the team looked deeper into the 5 systems with trends, they found that 4 of those trends were not due to giant planets. Two of the trends were due to stellar companions and two other trends were due to activity cycles of the stars, like our Sun’s 11 year activity cycle. That leaves the total stats for the survey of 4 giant planets found in 38 systems. Does this mean that the occurrence rate of giant planets in systems with inner sub-Neptunes is (4/38)*100% = 10.5%? Not quite. First the team had to determine their survey’s completeness. 

To determine how complete their survey was, the team wanted to characterize how sensitive their survey was to finding these kinds of distant giant planets in the first place. Because no survey can sample every star every night, no survey is fully complete. No team is ever awarded so much telescope time in the first place, but also things like weather and the seasons play a role in data collection, as well as the performance of the instrument. All of these things are working against planet detection. So, to determine their overall sensitivity, the team performed an injection/recovery analysis. Essentially, they take the real timestamps and error bars but simulate (inject) the velocities that a true giant planet would have imparted into their data. Then they see if they can confidently recover the signal using the same analysis techniques that they used on the real data. They run this injection many thousands of times for many different simulated planet parameters, each time trying to recover the signal. Based on how often they are correctly recovering the known injected signals, they can determine the completeness of the survey. 

Figure 2 (Figure 1 in the text): A completeness map of one star from the survey.  The y-axis represents an injected planet’s mass and the x-axis represents the injected planet’s semi-major axis, a proxy for orbital period. The white regions of parameter space are those where the survey was very sensitive to recovering planets. The dark regions are those where the survey was not sensitive to planets with those parameters. 

Using this completeness and the known number of planets found, the team determined that the occurrence rate of Jupiter-like planets in systems with known small planets is 9.3 +7.7 -2.9 %. This number is a bit different from occurrence rates found by other teams in similar studies, where slightly higher rates have been found. But the authors note that differences in how survey’s were carried out may contribute to these differences. For example, different survey’s had different criteria for what is considered a detection. Additionally, this new occurrence rate is consistent with the theory that there should be fewer Jupiter-like planets in systems with known sub-Neptunes, due to the possible hindrance effects that Jupiters could play when sub-Neptunes migrate from outer planetary systems to the inner part. Overall, this study sheds a bit more light on the question of: do Jupiters hurt small planet occurrences? And, how unique is our Solar System?

Astrobite edited by Roel Lefever

Featured image credit: Jack Lubin

About Jack Lubin

Jack received his PhD in astrophysics from UC Irvine and is now a postdoc at UCLA. His research focuses on exoplanet detection and characterization, primarily using the Radial Velocity method. He enjoys communicating science and encourages everyone to be an observer of the world around them.

Discover more from astrobites

Subscribe to get the latest posts to your email.

Leave a Reply