The Frequency of Low-Mass Exoplanets. II. The `Period Valley'

TL;DR

This study uses high-precision radial velocity data and two detection methods to analyze the frequency of low-mass exoplanets, confirming the existence of a period valley for giant planets and suggesting observational biases for intermediate-mass planets.

Contribution

It introduces two conservative, quantitative detection methods for low-mass exoplanets and applies them to establish the reality of the period valley and assess detection limits.

Findings

The period valley for giant planets (10-100 days) is confirmed as real.

Detection limits reach down to 3 m/s velocity amplitude for the best stars.

The apparent deficit of 10-100 M_earth planets may be due to selection effects.

Abstract

Radial-velocity planet search campaigns are now beginning to detect low-mass "Super-Earth" planets, with minimum masses M sin i < 10 M_earth. Using two independently-developed methods, we have derived detection limits from nearly four years of the highest-precision data on 24 bright, stable stars from the Anglo-Australian Planet Search. Both methods are more conservative than a human analysing an individual observed data set, as is demonstrated by the fact that both techniques would detect the radial velocity signals announced as exoplanets for the 61 Vir system in 50% of trials. There are modest differences between the methods which can be recognised as arising from particular criteria that they adopt. What both processes deliver is a quantitative selection process such that one can use them to draw quantitative conclusions about planetary frequency and orbital parameter distribution from a given data set. Averaging over all 24 stars, in the period range P<300 days and the eccentricity range 0.0<e<0.6, we could have detected 99% of planets with velocity amplitudes K>7.1 m/s. For the best stars in the sample, we are able to detect or exclude planets with K>3 m/s, corresponding to minimum masses of 8 M_earth (P=5 days) or 17 M_earth (P=50 days). Our results indicate that the observed "period valley," a lack of giant planets (M>100 M_earth) with periods between 10-100 days, is indeed real. However, for planets in the mass range 10-100 M_earth, our results suggest that the deficit of such planets may be a result of selection effects.