A Jovian analogue orbiting a white dwarf star.

Studies1,2 have shown that the remnants of destroyed planets and debris-disk planetesimals can survive the volatile evolution of their host stars into white dwarfs3,4, but few intact planetary bodies around white dwarfs have been detected5-8. Simulations predict9-11 that planets in Jupiter-like orbits around stars of ≲8 M☉ (solar mass) avoid being destroyed by the strong tidal forces of their stellar host, but as yet, there has been no observational confirmation of such a survivor. Here we report the non-detection of a main-sequence lens star in the microlensing event MOA-2010-BLG-477Lb12 using near-infrared observations from the Keck Observatory. We determine that this system contains a 0.53 ± 0.11 M☉ white-dwarf host orbited by a 1.4 ± 0.3 Jupiter-mass planet with a separation on the plane of the sky of 2.8 ± 0.5 astronomical units, which implies a semi-major axis larger than this. This system is evidence that planets around white dwarfs can survive the giant and asymptotic giant phases of their host's evolution, and supports the prediction that more than half of white dwarfs have Jovian planetary companions13. Located at approximately 2.0 kiloparsecs towards the centre of our Galaxy, it is likely to represent an analogue to the end stages of the Sun and Jupiter in our own Solar System.

The frequency of snowline-region planets from four-years of OGLE-MOA-Wise second-generation microlensing.

We present a statistical analysis of the first four seasons from a "second-generation" microlensing survey for extrasolar planets, consisting of near-continuous time coverage of 8 deg2 of the Galactic bulge by the OGLE, MOA, and Wise microlensing surveys. During this period, 224 microlensing events were observed by all three groups. Over 12% of the events showed a deviation from single-lens microlensing, and for ~1/3 of those the anomaly is likely caused by a planetary companion. For each of the 224 events we have performed numerical ray-tracing simulations to calculate the detection efficiency of possible companions as a function of companion-to-host mass ratio and separation. Accounting for the detection efficiency, we find that 55 - 22 + 34 % of microlensed stars host a snowline planet. Moreover, we find that Neptunes-mass planets are ~ 10 times more common than Jupiter-mass planets. The companion-to-host mass ratio distribution shows a deficit at q ~ 10-2, separating the distribution into two companion populations, analogous to the stellar-companion and planet populations, seen in radial-velocity surveys around solar-like stars. Our survey, however, which probes mainly lower-mass stars, suggests a minimum in the distribution in the super-Jupiter mass range, and a relatively high occurrence of brown-dwarf companions.

Exoplanet detection. A terrestrial planet in a ~1-AU orbit around one member of a ~15-AU binary.

Using gravitational microlensing, we detected a cold terrestrial planet orbiting one member of a binary star system. The planet has low mass (twice Earth's) and lies projected at ~0.8 astronomical units (AU) from its host star, about the distance between Earth and the Sun. However, the planet's temperature is much lower, <60 Kelvin, because the host star is only 0.10 to 0.15 solar masses and therefore more than 400 times less luminous than the Sun. The host itself orbits a slightly more massive companion with projected separation of 10 to 15 AU. This detection is consistent with such systems being very common. Straightforward modification of current microlensing search strategies could increase sensitivity to planets in binary systems. With more detections, such binary-star planetary systems could constrain models of planet formation and evolution.