Thursday, October 23, 2014

An Ancient Relic of the Universe’s First Galaxies

Ultra-faint dwarf galaxies (UFDs) are the faintest galaxies known in the universe. These galaxies are smaller than 1,000 light years in radius and have very low metallicities. As a result, UFDs are believed to represent the first generation of galaxies in the universe. UFDs are the most dark matter dominated galaxies since nearly all their mass is admittedly in the form of dark matter. In fact, the total mass of all stars in a UFD is typically less than a million solar masses. For comparison, the Milky Way galaxy contains 200 to 400 billion stars. UFDs are also much smaller and fainter than classical dwarf spheroidal galaxies. So far, UFDs have only been discovered around the Milky Way galaxy and the neighbouring Andromeda galaxy.


Using images acquired by the Hubble Space Telescope, Jang & Lee (2014) reported the discovery of a UFD in a region of space far from any massive galaxy in the Virgo Cluster. This newly discovered UFD is named Virgo UFD1. Observations indicate that Virgo UFD1 contains low metallicity red-giant-branch (RGB) stars but no asymptotic-giant-branch (AGB) stars. It means that Virgo UFD1 is very old, most likely older than 10 billion years, because AGB stars mark the final evolution of stars that are at least as massive as the Sun and such stars do not live more than about 10 billion years.

The stars in Virgo UFD1 probably formed in a single burst of star formation more than 10 billion years ago. As a consequence, the absence of subsequent generations of stars to fuse hydrogen and helium into heavier elements results in the low metallicity of Virgo UFD1. There is also no sign that Virgo UFD1 experienced tidal interaction with any galaxy. Estimates place Virgo UFD1 at a distance of more than 50 million light years away. The old age, low metallicity and large distance from any massive galaxy suggest Virgo UFD1 may represent an ancient relic of the universe’s first galaxies.

Left panel: Effective radius versus absolute total magnitude of Virgo UFD1 (large starlet symbol) in comparison with other stellar systems. Right panel: The central surface brightness versus absolute total magnitude of Virgo UFD1 in comparison with other stellar systems. Note: Circles and lenticular symbols for the giant ellipticals and bulges in spiral galaxies, downward triangles for the UCDs, pentagons for the Milky Way galaxy globular clusters, squares and diamonds for the Local Group satellite galaxies and UFDs, and upward triangles for the dwarf galaxies in M81 and M106 and the low surface brightness galaxies in M101, and small yellow starlet for Virgo dSph-D07. Jang & Lee (2014).

Reference:
Jang & Lee (2014), “Discovery of an Ultra-Faint Dwarf Galaxy in the Intracluster Field of the Virgo Center: A fossil of the First Galaxies”, arXiv:1410.2247 [astro-ph.GA]

Tuesday, October 14, 2014

A Warm Gas-Giant Planet Orbiting a Giant Star

A study by Ortiz et al. (2014) reported on the spectroscopic confirmation of a Jupiter-like gas-giant planet in a close-in, eccentric orbit around a giant star. This planet is identified as KOI-1299b, with the suffix “b” denoting its planetary nature. The host star of KOI-1299b is entering its later stages of stellar evolution and is about to swell into a red giant star. This star has 1.35 ± 0.10 times the Sun’s mass, 4.15 ± 0.12 times the Sun’s radius and an effective surface temperature of 5020 ± 60 K. KOI-1299b is observed by NASA’s Kepler space telescope to transit its host star once every 52.5 days. The size of KOI-1299b was estimated by measuring the drop in the star’s brightness each time the planet transits in front of it. A separate paper by Ciceri et al. (2014) on the discovery of KOI-1299b also reported similar results.

Figure 1: Artist’s impression of a gas-giant planet with a system of planetary rings circling it.

The planetary nature of KOI-1299b was confirmed using high-resolution spectroscopic follow-up observations conducted between June and October 2014 using the Calar Alto Fiber-fed Échelle spectrograph (CAFE) on the 2.2 m telescope of Calar Alto Observatory in Almería, Spain, and the Fibre-fed Échelle Spectrograph (FIES) on the 2.56 m Nordic Optical Telescope of Roque de los Muchachos Observatory in La Palma, Spain. These follow-up observations measured the amount of gravitational tugging KOI-1299b exerts on its host star, allowing the mass of KOI-1299b to be estimated and its planetary nature to be confirmed. Additionally, these follow-up observations also show that KOI-1299b has a highly eccentric orbit around its host star.

Figure 2: Radial velocity measurements of the host star of KOI-1299b. Upper panel: CAFE (blue circles) and FIES (red squares). Lower panel: Residuals. Ortiz et al. (2014)

 Figure 3: Left panel: Eccentricity and semimajor axis of the extrasolar planets discovered around main sequence stars (black dots) and giant stars (magenta circles). Right panel: Orbital period versus stellar mass. The position of KOI-1299b is marked with a green triangle in both panels. Ortiz et al. (2014)

KOI-1299b has 5.86 ± 0.05 times the mass of Jupiter, 1.08 ± 0.03 times the radius of Jupiter and an orbital eccentricity of 0.479 ± 0.004. With the mass and size known, the estimated density of KOI-1299b is 5.7 ± 0.5 g/cm³. The highly eccentric orbit of KOI-1299b brings the planet as close as ~0.16 AU (periastron) from its host star and out as far as ~0.45 AU (apastron). Between periastron and apastron, KOI-1299b receives ~450 to ~56 times as much insolation as Earth receives from the Sun. On average, the estimated equilibrium temperature of KOI-1299b is 942 ± 20 K. However, the planet’s highly eccentric orbit can cause temperatures to vary by ~500 K. KOI-1299b is classed as a warm-Jupiter since is not as hot as typical hot-Jupiters whose temperatures are well over 1000 K. As the host star of KOI-1299b swells into a red giant, tidal interactions between the star and planet will increase, eventually causing KOI-1299b to be engulfed by its host star.

References:
- Ortiz et al. (2014), “Spectroscopic confirmation of KOI-1299b: a massive warm Jupiter in a 52-day eccentric orbit transiting a giant star”, arXiv:1410.3000 [astro-ph.EP]
- Ciceri et al. (2014), “KOI-1299b: a massive planet in a highly eccentric orbit transiting a red giant”, arXiv:1410.2999 [astro-ph.EP]

Thursday, October 9, 2014

Polluting a Red Supergiant Star with Heavy Elements

A recent study by Levesque et al. (2014) found that the red supergiant star HV 2112 in a nearby galaxy known as the Small Magellanic Cloud is enriched with various peculiar heavy elements. As a result, HV 2112 is postulated to be a Throne-Zytkow Object (TZO). Basically, a TZO is a red supergiant star with a neutron star at its center. The neutron star most likely got there by in-spiralling within the envelope of the red supergiant star, all the way down to the core. The neutron star destroys the red supergiant star’s core and part of the core forms an accretion disk around the neutron star. Temperatures and densities in the accretion disk are high enough to synthesize a range of peculiar heavy elements.



HV 2112 is observed to be enriched with heavy elements such as calcium, rubidium, lithium and molybdenum. Although most of these heavy elements can be synthesized by a red supergiant star on its own, the high calcium abundance requires HV 2112 to be a TZO since only in an accretion disk around a neutron star are the temperatures and densities sufficiently high to synthesize calcium. Nevertheless, a study by Sabach & Soker (2014) provides an alternative explanation whereby the high calcium abundance observed for HV 2112 came from a companion star that had exploded as a core collapse supernova (CCSN). During the supernova event, HV 2112 was already a large red supergiant star and could intercept more of the supernova’s ejecta, including the calcium that was synthesized in the supernova.

HV 2112 and its companion star started of as a pair of massive stars circling one another. The companion star is slightly more massive than HV 2112. Since more massive stars evolve quicker, the companion star first evolves to a red giant star, puffs up, and transfers some of its mass to HV 2112. Eventually, HV 2112 becomes more massive than its companion star and also evolves to become a red giant star. The companion star subsequently explodes as a CCSN. When that happens, HV 2112 has already evolved further to form a red supergiant star. Its large size allows HV 2112 to intercept a good fraction of the supernova ejecta that is rich in heavy elements, including calcium. As a result, HV 2112 could be a red supergiant star that was “polluted” by ejecta from a supernova, rather than a TZO.

References:
- Levesque et al. (2014), “Discovery of a Thorne-Zytkow object candidate in the Small Magellanic Cloud”, arXiv:1406.0001 [astro-ph.SR]
- Sabach & Soker (2014), “A super asymptotic giant branch star enriched with calcium by a supernova as the origin of HV2112, rather than a Thorne-Zytkow Object”, arXiv:1410.1713 [astro-ph.SR]

Wednesday, October 8, 2014

One Planet, Two Stars

Welsh et al. (2014) present the discovery of KIC 9832895b, a circumbinary planet in a 240.5 day orbit around an eclipsing binary. Basically, an eclipsing binary is a pair of stars that appear to eclipse one another as they orbit around each other. In the case of KIC 9832895b, it orbits around an eclipsing binary consisting of a pair of stars with 0.93 and 0.194 times the Sun’s mass. Both stars orbit around one another with a period of 27.3 days. The orbital period of KIC 9832895b is 8.8 times the orbital period of the eclipsing binary. This places the planet safely outside the dynamical instability zone.

Figure 1: Artist’s impression of a gaseous planet which KIC 9832895b might resemble.

Figure 2: Face-on view of KIC 9632895b’s orbit, showing the habitable zone (HZ). The dark green region corresponds to the narrow (conservative) HZ and the light green corresponds to the nominal (extended) HZ. The dashed red circle represents the dynamical instability zone. The orbit of KIC 9632895b is shown in white. Welsh et al. (2014).

KIC 9832895b was detected from its three transits across the primary star (i.e. the more mass star) of the eclipsing binary. The transit depth indicates that KIC 9832895b is 6.2 times the radius of Earth, indicating that KIC 9832895b is somewhat larger than Neptune. The mass of KIC 9832895b is estimated to be most likely less than 16 times the mass of Earth due to the absence of any noticeable perturbations it has on the eclipsing binary. This constrains the mean density of KIC 9832895b to be less than 0.38 g/cm³ and demonstrates that it is an unusually low density planet, probably of gaseous composition.

KIC 9832895b is the 10th circumbinary planet discovered using data collected from NASA’s Kepler space telescope. In addition, KIC 9832895b is also in the circumbinary habitable zone where temperatures are relatively clement. The time-averaged insolation that KIC 9832895b receives is estimated to be 94 percent the intensity of insolation Earth receives from the Sun. Although KIC 9832895b is itself unlikely to harbour life, it could host a large moon capable of supporting life. Of the 10 circumbinary planets known so far, KIC 9832895b is the third found to lie within the circumbinary habitable zone.

Interestingly, the inclination of KIC 9832895b oscillates with a 102.8 year period. As a result, transits only occur ~8 percent of the time. This explains why the three detected transits of KIC 9832895b were only found in the later portion of the Kepler dataset. The transits will not be observable after 2015 and will only return on 2066. Since the transits do not always happen, for every system like the one hosting KIC 9832895b, there are ~12 similar systems where planetary transits are not observed.

Reference:
Welsh et al. (2014), “KIC 9632895 - The 10th Kepler Transiting Circumbinary Planet”, arXiv:1409.1605 [astro-ph.EP]

Sunday, October 5, 2014

Identifying Alien Planets with Clear Skies

When a planet transits in front of its host star, a tiny fraction of the starlight passes through the planet’s atmosphere and carries with it signatures of the planet’s atmospheric constituents. This can allow the planet’s atmosphere to be characterised using an observational technique known as transmission spectroscopy. However, the atmospheres of planets can be cloudy, hazy or clear-sky (i.e. free of clouds and hazes). The presence of clouds or hazes can obscure the lower layers of the atmosphere and make the planet less desirable for characterisation. As a result, it is worth identifying whether a planet has clear skies before a large amount of telescope time is dedicated to characterising its atmosphere.


Misra & Meadows (2014) propose a method to readily distinguish cloudy, hazy and clear-sky planets. This involves measuring the amount of starlight being refracted through the atmospheres of transiting planets using upcoming large collecting area space and ground-based telescopes such as the James Webb Space Telescope (JWST) and the European Extremely Large Telescope (E-ELT). The refraction of starlight by a planet’s atmosphere can lead to an increase of flux both prior to ingress (i.e. before the start of a transit) and subsequent to egress (i.e. after the end of a transit).

The presence of a global cloud or haze coverage tends to obscure layers of a planet’s atmosphere that refract light. As a result, the detection of refracted light pre-ingress and post-egress would strongly suggest the absence of a global cloud or haze layer, making the planet a promising candidate for follow-up observations to characterise its atmosphere. In the models, the atmospheric pressure cut-offs are at 1 mbar (hazy case), 0.1 bars (cloudy case) and 1 bar (clear-sky case). A higher pressure cut-off indicates a greater depth of measurable atmosphere.

Results from the study show detecting refracted light requires less than 10 hours of total observing time for Jupiter-sized planets with JWST and for Super-Earths/Mini-Neptunes with E-ELT. Since the increase in flux due to refraction prior to ingress and subsequent to egress can be readily detected for clear-sky planets, it can quickly identify whether a planet is a good candidate for extended follow-up observations. Characterising a planet’s atmosphere is a very time consuming process, making it important to select good candidates (i.e. clear-sky planets) prior to characterisation.

Reference:
Misra & Meadows (2014), “Discriminating Between Cloudy, Hazy and Clearsky Exoplanets Using Refraction”, arXiv:1409.7072 [astro-ph.EP]

Friday, October 3, 2014

An Oblate Giant Planet

Kepler-39b is a gas-giant planet in orbit around an F-type star. It is 18 times Jupiter’s mass, 1.22 times Jupiter’s radius and it transits its host star every 21.09 days. A study by Wei Zhu et al. (2014) using data from NASA’s Kepler space telescope found that Kepler-39b has an oblateness of 0.22 ± 0.11. In fact, this is the first tentative detection of oblateness for a planet outside the Solar System. When an oblate planet transits its host star, the transit light curve will exhibit small differences from that of a purely spherical planet.  


In the Solar System, the gas-giant planets Jupiter and Saturn are oblate in shape due to their rapid rotations. The oblateness of an object is expressed as the ratio of its equatorial-polar radius difference to its equatorial radius. The equatorial radius is larger than the polar radius by 7 percent for Jupiter and by 10 percent for Saturn. As such, Jupiter’s oblateness is 0.07 and Saturn’s oblateness is 0.1. With an oblateness of 0.22 ± 0.11, Kepler-39b is substantially more oblate than any planet in the Solar System.

The large oblateness of Kepler-39b is most likely rotationally induced. With that, its rotation period is estimated to be 1.6 ± 0.4 hours. For comparison, the rotation periods of Jupiter and Saturn are 9.9 and 10.6 hours, respectively. Although the rotation of Kepler-39b is remarkably fast, it is lower than its estimated break-up rotation period of ~0.9 hours. In addition to its large oblateness, Kepler-39b is also inflated in size. Its close proximity to its host star and its estimated equilibrium temperature of around 900 K is insufficient to account for its inflated size.

Reference:
Wei Zhu et al. (2014), “Constraining the Oblateness of Kepler Planets”, arXiv:1410.0361 [astro-ph.EP]

Wednesday, October 1, 2014

When Supermassive Stars Explode in the Early Universe

Supermassive stars with ~10,000 to ~100,000 times the Sun’s mass are believed to have formed in the very early universe. These are the first generation of stars in the universe and are entirely comprised of hydrogen and helium. They live very short lives before collapsing directly to form black holes. A team of astrophysicists ran a number of supercomputer simulations and found that some of these supermassive stars die in a rather unusual way. Instead of collapsing to form black holes, supermassive stars in a narrow mass range between 55,000 to 56,000 times the Sun’s mass explode as highly energetic thermonuclear supernovae, leaving nothing behind.

This image is a slice through the interior of a supermassive star of 55,500 solar masses along the axis of symmetry. It shows the inner helium core in which nuclear burning is converting helium to oxygen, powering various fluid instabilities (swirling lines). This snapshot shows a moment one day after the onset of the explosion, when the radius of the outer circle would be slightly larger than that of the orbit of the Earth around the Sun. (Credit: Ken Chen, UC Santa Cruz)

A supermassive star with 55,500 times the Sun’s mass lives for about 1.69 million years before it becomes unstable and starts to collapse. During its pre-collapse phase, the size of the star is slightly larger than the diameter of Earth’s orbit around the Sun and the star has an effective surface temperature of about 70,000 K. The star is also remarkably luminous, with ~1.5 billion times the Sun’s luminosity. With the onset of helium burning in the star’s core, the prodigious amount of thermal photons being generated in the core affects the star’s gravitational field by becoming an additional source of gravity. As a consequence, the core begins to contract, causing the temperature and density in the core to rise rapidly, accelerating nuclear burning.

As the core contracts, helium begins to burn explosively, fusing to carbon, and then to oxygen, neon, magnesium and silicon. The explosive nuclear burning occurs within a span of only several hours and releases ~10 times more energy than the binding energy of the star. This causes the star to halt its collapse and unbind completely in a massive explosion known as a general relativistic supernova (GSN). The amount of energy produced in such an event is ~10,000 times the energy released by a typical supernova. In total, about half the mass of the star is ejected in the form of elements heavier than hydrogen and helium. Mostly elements between carbon and silicon are produced, with only trance amounts of iron group elements.

After getting blasted out into the cosmos, these heavy elements are incorporated in the formation of subsequent generations of stars and planets. The energetic demise of these supermassive stars can be detected by upcoming space-based observatories such as ESA’s Euclid and NASA’s Wide-Field Infrared Survey Telescope (WFIRST). Additionally, indirect observational signatures of GSN explosions might be found by looking for early galaxies that are iron deficient but enhanced with elements from carbon to silicon.

Reference:
Ke-Jung Chen et al. (2014), “General Relativistic Instability Supernova of a Supermassive Population III Star”, arXiv:1402.4777 [astro-ph.HE]

Tuesday, September 30, 2014

Hot Giant Planet that is Blacker than Coal

Gandolfi et al. (2014) report on the discovery of a half-Jupiter mass planet transiting an old Sun-like star every 2.7 days. This discovery combines data collected by NASA’s Kepler space telescope from 13 May 2009 to 11 May 2013 with spectroscopic follow-up observations performed with the FIES spectrograph at the Nordic Optical Telescope in La Palma, Spain. Photometric data from Kepler indicates how much starlight is blocked when the planet transits in front of its host star, allowing the size of the planet to be estimated. The FIES spectrograph measures the amount of gravitational tugging the planet has on its host star and provides the estimated mass of the planet.

Figure 1: Artists’ illustration of a hot-Jupiter orbiting a Sun-like star. Image credit: Haven Giguere & Nikku Madhusudhan.

Figure 2: Phase-folded transit light curve of KOI-183b showing the best fitting model and residuals. Gandolfi et al. (2014).

Figure 3: Radial velocity data from the FIES spectrograph with the median, 68th and 99th percentile limits. Gandolfi et al. (2014).

The planet, identified as KOI-183b, is estimated to have 0.595 ± 0.081 times the mass of Jupiter and 1.192 ± 0.052 times the radius of Jupiter. Given the mass and size of the planet, its bulk density is 0.459 ± 0.083 g/cm³. KOI-183b orbits its host star at a distance of only ~1/28th the Earth-Sun distance. As a result, KOI-183b is intensely heated and is classified as a hot-Jupiter. The radius of KOI-183b is consistent with theoretical models for heavily irradiated coreless gas-giant planets. Being so near to its host star, temperatures on KOI-183b can reach ~2000 K, hot enough to melt titanium metal.

Data from Kepler also indicates that KOI-183b periodically passes behind its host star in what is known as a secondary eclipse. The secondary eclipse signal has a depth of 14.2 ± 6.6 ppm. From the depth of its secondary eclipse signal, KOI-183b is estimated to have a very low Bond albedo of only 0.037 ± 0.019, making it one of the “darkest” gas-giant planets known so far. Basically, KOI-183b reflects only ~4 percent of the incoming radiation from its host star back into space. For comparison, that is darker than coal. Other hot-Jupiters with similarly low Bond albedos include TrES-2b and Kepler-77b.

Reference:
Gandolfi et al. (2014), “KOI-183b: a half-Jupiter mass planet transiting a very old solar-like star”, arXiv:1409.8245 [astro-ph.EP]

Monday, September 29, 2014

A Highly Eccentric Brown Dwarf around a Giant Star

To date, ~10 brown dwarfs are known around giant stars (i.e. evolved stars). Brown dwarf are objects more massive than planets, but are not massive enough to count as full-fledged stars. M. I. Jones et al. (2014) report on the discovery of a brown dwarf on a highly eccentric orbit around the giant star HIP 97233. The brown dwarf, identified as HIP 97233 b, has an orbital period of 1058.8 days and a minimum mass of 20 times the mass of Jupiter.

With an orbital eccentricity of 0.61, HIP 97233 b is the brown dwarf with the most eccentric orbit known around a giant star. The mass and orbit of HIP 97233 b were both determined from the gravitational “tugging” it exerts on its host star which was observed in the form of a radial velocity signature (i.e. Doppler shifts in the star’s spectral lines).

Figure 1: Artist’s impression of a giant planet.

 Figure 2: Upper panel: Radial velocity curve for the host star of HIP 97233 b. Lower panel: Residuals from the best fit. M. I. Jones et al. (2014).

HIP 97233 b highly eccentric orbit takes it from as near as 1.0 AU to as far as 4.1 AU from its host star. M. I. Jones et al. (2014) estimate that the host star of HIP 97233 b has 1.84 ± 0.14 times the Sun’s mass and 5.20 ± 0.50 times the Sun’s radius. The host star of HIP 97233 b is considerably larger and more luminous than the Sun. At closest approach, the dayside of HIP 97233 b receives roughly 16 times the intensity of insolation as Earth receives from the Sun.

There are a number of ways through which an object like HIP 97233 b can form. Firstly, the host star of HIP 97233 b is much more massive than the Sun, enabling it to have a more massive protoplanetary disk which can allow massive planets and brown dwarfs to form more efficiently. Also, as the star evolves and swells in size, it begins to blow an enhanced stellar wind from which a giant planet can accrete a significant amount of mass and grow in mass till it reaches the brown dwarf mass regime.

The star’s high metallicity might also have enabled HIP 97233 b to form by core accretion, believed to be the main mechanism through which planets form. Finally, interaction with the protoplanetary disk before it was dissipated or with the star’s outer layers as it evolves to a giant star might have caused HIP 97233 b to migrate inward from beyond ~4 AU to where it currently is.

Reference:
M. I. Jones et al. (2014), “A planetary system and a highly eccentric brown dwarf around the giant stars HIP 67851 and HIP 97233”, arXiv:1409.7429 [astro-ph.EP]

Wednesday, August 27, 2014

A Uranus-Type Planet in a Binary Stellar System

A gravitational microlensing search by R. Poleski (2014) revealed the presence of a Uranus-type planet in orbit around a 0.6 solar mass star. The gravitational microlensing event is designated OGLE-2008-BLG-092, and the newfound planet is estimated to be ~3 times the mass of Uranus and it circles its host star at ~16 AU. For comparison, Uranus orbits the Sun at an average distance of 19 AU. This newfound planet is the first known exoplanet whose mass and orbit is similar to Uranus. The planet was detected when it and its host star fortuitously passed in front of a background star, and the gravitational field of the star-planet system magnified light from the background star.

Figure 1: Artist’s impression of a Uranus-type planet.

Planets in the Solar System can be classed into 3 groups: small rocky planets (Earth, Venus, etc), gas giants (Jupiter and Saturn) and ice-giants (Uranus and Neptune). At present, the leading methods of detecting planets around other stars (i.e. transit and radial velocity methods) have yet to turn up any extrasolar analogues of Uranus and Neptune. Such planets are far from their host stars and have orbital periods that exceed a human lifespan. As a result, both the transit and radial velocity methods have yet to turn up such planets since both methods greatly favour the detection of planets with short orbital periods. To detect extrasolar analogues of Uranus and Neptune using such methods would require exceedingly long observation timescales.

Although the technique of direct imagine can detect planets that orbit far from their host stars, this technique has so far been restricted to the detection of more massive and hotter planets that inhibit young planetary systems. These planets are very different from planets like Uranus and Neptune. At present, the only method that can detect extrasolar analogues of Uranus and Neptune seems to be gravitational microlensing as this method allows planets to be detected regardless of their orbital periods. In addition to the Uranus-type planet and its host star, the OGLE-2008-BLG-092 microlensing event also revealed the presence of a companion object in the system that is either a low mass star or a brown dwarf. In fact, the projected separation of the Uranus-type planet from its host star is only ~3 times smaller than that of the companion star (or brown dwarf).

Figure 2: Light curve of the OGLE-2008-BLG-092 microlensing event. The inset shows the planetary subevent. The presence of the companion star (or brown dwarf) is indicated by the 2010 subevent. R. Poleski (2014).

Reference:
R. Poleski (2014), “Triple Microlens OGLE-2008-BLG-092L: Binary Stellar System with a Circumprimary Uranus-type Planet”, arXiv:1408.6223 [astro-ph.EP]