Anybody out there?

Since the mid-20th century scientists have been searching for potentially habitable extrasolar planets, or exoplanets. With the introduction of space-based observation platforms such as NASA’s Kepler Mission, the number of exoplanets that have been found is growing.

Figure 1a: The percentages indicate the increase in the numbers of respective size category planet candidates since the last Kepler catalogue was released, in February 2012. To identify the potential planet candidates, more than 13,000 transit-like signals, called 'threshold crossing events', were analysed to eliminate known spacecraft instrumentation and astrophysical false positives, phenomena that masquerade as planetary candidates.

“ARE we alone in the universe?” is a question that humankind has been asking for millennia. Within the solar system itself, conditions on Mars and Europa, one of Jupiter’s moons, could favour some form of life. Scientists generally believe that it is highly plausible that there is life in other planetary systems in the galaxy, but there is no evidence of it yet. In fact, on the basis of the hypothesis that biological complexity increased exponentially during evolution, Alexei Sharov of the National Institutes of Health of the United States concluded in a 2006 study that life in the universe may have begun 10 billion years ago, about six billion years before the earth came into existence.

Different kinds of searches for indications of extraterrestrial life have been going on since the mid-20th century, from detecting radio signals from distant intelligent life using ground-based radio astronomy in the SETI (Search for Extraterrestrial Intelligence) project to searching for potentially habitable extrasolar planets, or exoplanets, using both ground-based and space-based astronomical techniques. This has of late resulted in news headlines such as “A new planet has been discovered orbiting star X at Y light years away”.

The first exoplanet was discovered, arguably, in 1988 by David W. Latham of the Harvard-Smithsonian Centre for Astrophysics (CfA) and others. Arguably because even to date there is only insufficient confirmation that HD 114762b, as it is called, is a true planet. It is still not clear whether HD 114762b, located 132 light years away, is a high-mass gas giant—a Jupiter-like planet that is not primarily composed of rock or other solid matter—or a brown dwarf or a red dwarf star. (The solar system has four gas giants: Jupiter, Saturn, Uranus and Neptune.) Secondly, also in 1988, a team of Canadian astronomers saw indications of a possible planet (Gamma Cephei Ab) around the star Gamma Cephei A 45 light years away and claimed so in a paper. But in 1992, the claim was withdrawn owing to insufficient evidence. Later, in 2002, its status as an exoplanet was reinstated on the basis of improved measurements.

The first clear detection of an exoplanet was in 1992 when Aleksander Wolszczan and Dale Frail found three planets orbiting the pulsar PSR1257+12, a rather unexpected environment. The first exoplanet around a “normal” (main sequence) star, 51 Pegasi, was discovered in 1995 by Michel Mayor and Didier Queloz. With vast improvements in planet-detecting techniques since then, particularly with the introduction of space-based observation platforms—the French space agency’s COROT (COnvection ROtation et Transits planétaires) mission, launched in December 2006, and the American National Aeronautics and Space Administration’s Kepler space telescope mission, launched in March 2009—the number of exoplanets that have been found is growing rapidly. Most of these are in the Milky Way, but a few extragalactic planets have also been detected. As of date, the total number of exoplanets discovered stands at 861 from 677 planetary systems, 128 of which are multi-planet. But, as of January 7, 2013, Kepler has identified as many as 2,740 exoplanet candidates, orbiting 2,036 stars, from its 22 months of data, gathered from May 2009 to March 2011. These are called Kepler Objects of Interest (KOIs) (Figure 1a). Of these, 114 are confirmed planets and the rest await confirmation. We shall return to what confirmation entails later.

Compared with the Kepler data released in February 2012, the increase in the number of candidates is 20 per cent. The most dramatic increase has been in the number of earth-size (1-1.25 times the radius of the earth radius, RE) and super-earth size (1.25-2 times RE) candidates (43 per cent and 21 per cent respectively; Figure 1b). The number of stars with more than one KOI is 467, and 43 per cent of KOIs have neighbouring planets. The large number of multi-candidate systems implies that a substantial fraction of exoplanets resides in flat multi-planet systems, which is consistent with our neighbourhood in the solar system.

It should be borne in mind that the Kepler Mission is not an all-sky search. It is designed to continuously and simultaneously monitor, over its 3.5-year lifetime, the brightnesses of about 150,000 stars brighter than 14th magnitude stars in the constellations Cygnus and Lyrae (Figure 2), which lie at distances of about a few hundred to a few thousand light years away. The Cygnus-Lyra region in the northern region was chosen for its rich field of stars, somewhat richer than a southern field. Also, since it is faraway north from the plane of the earth’s orbit, the sun does not get in the way of observations throughout the entire orbit of the Kepler spacecraft. The Kepler satellite has a 0.95-metre-diameter telescope, which is a photometer—a sensitive light meter—with a field of view (FOV) of about 105 square degrees (which is an area of sky the size of about two open hands). The FOV of most telescopes is less than one square degree. Any all-sky inference from Kepler data is, therefore, a statistical extrapolation of this limited data from a particular patch in the northern sky.

According to one such CfA analysis presented at the January meeting of the American Astronomical Society, the fraction of stars that have an earth-sized planet (0.8-1.25 RE) in an orbit closer than Mercury, and an orbit period of 85 days or less, is 17 per cent. About 20 per cent have a super-earth in an orbit of 150 days or less. The same fraction has a mini Neptune-like planet with up to a 250-day orbit. This shows that more than half the stars in the galaxy have an earth-size or larger planet in a close orbit, and if one also includes larger planets with wider orbits up to the orbital distance of the earth (one astronomical unit, or AU), the fraction goes up to 70 per cent. The analysis also found that the type of star, whether it is a sun-like star or a red dwarf, does not really matter for all planet sizes except for gas giants. This is an important finding towards the understanding of the formation of planetary systems.

Habitable zone

Since the Milky Way has about 100 billion stars, the above analysis suggests that there are at least about 17 billion earth-sized planets in the galaxy. However, all of them may not be habitable. Of particular interest, however, is the existence of terrestrial exoplanets that can support life. That is, to identify earth-like planets that lie, just like the earth, in the “habitable zone”, or the “Goldilocks zone” as it also called, around the host star (at 0.8-1.6 AU distance), where the conditions are just right to sustain life. Since we know that liquid water is essential for all forms of life, the habitable zone is where it is theoretically possible for a planet with sufficient atmospheric pressure to maintain liquid water on its surface.

Before the advent of Kepler, there were only a few exoplanets considered to be prime candidates for some form of life to exist on them, such as the super-earth-sized Gliese 581 c, g and d around the star Gliese 581, and OGLE-2005-BLG-390Lb, which were detected in the habitable zone using ground-based techniques. One of the prime objects of Kepler is to seek earth-size planets in the habitable zones of sun-like stars. In the latest Kepler update, several new habitable candidates have been found (Figure 3), of which four are less than 2 RE. In fact, one of these, designated as KOI-172.02, is Kepler’s first habitable-zone super-earth candidate around a sun-like star. The planet candidate is about 1.5 times RE, orbits the host star every 242 days, and is at a distance about 0.75 AU from the host star. Scientists believe that it is most likely to have water. On the basis of the findings of Kepler and other experiments, it is believed that 3-10 per cent of earth-sized and super-earth exoplanets could be habitable.

Searching for exoplanets is like looking for a needle in a haystack. Planets emit practically no light of their own. So detecting an exoplanet through direct imaging is very difficult because the faint light from the orbiting planet will be lost in the extreme glare of the light from the host star. Thus, to obtain the planet’s image, the light from the star has to be dimmed or masked. Therefore, the majority of exoplanets have been discovered through indirect detection only, which is based on the different effects a planet has on its host star. Basically, five different techniques are used for indirect detection of exoplanets: radial velocity (RV) measurements, planet’s transit in front of the host star, astrometry, pulsar timing, and gravitational microlensing. The first two are the chief methods used and, indeed, the largest number found so far has been with these two techniques.

The RV of a star is the velocity of the star along the line of sight of an observer on the earth. The RV tracking technique for detecting planets depends on the weak gravitational tug of the planet on the star. As the planet moves in its orbit, its gravitational pull makes the star execute a small circular orbit, introducing a minute wobble in its motion that can be detected. As the star moves in this orbit, it will move towards and away from the earth periodically by an amount that depends on the planet’s mass and the inclination of its orbit.

Changes in the RV in response to the gravitational pull of the planet will cause the wavelengths of the star’s spectral lines to be shifted towards the red when the star is moving away from the observer and towards the blue when it is approaching the observer. This is the Doppler effect (Figure 4) familiar to us from sound waves; for example, the change in the pitch of a police siren when it passes us on a street. Using high-precision spectrographs, these Doppler shifts in the star’s spectrum can be measured to estimate the small variations in the star’s RV. However, since the inclination of the planetary orbit is unknown, the measurement of this periodic variation yields only a minimum mass value for the planet.

At present, the most successful experiment to detect low-mass exoplanets using the RV method is HARPS (High-Accuracy Radial Velocity Planetary Searcher), which was mounted in 2002 on the European Southern Observatory’s (ESO) 3.6-m telescope at La Silla, Chile. Although designed to reach a precision level of detecting only 1 metre/second variation in RV, it is stated to be doing better than that, achieving about 0.5 m/s. The variation in the sun’s RV due to the earth’s pull is about 9 centimetres/s, which is slower than our walking speed. An improved spectrograph, ESPRESSO (Echelle SPectrograph for Rocky Exoplanet and Stable Spectroscopic Observations) is the proposed new-generation high-precision instrument to be mounted on the ESO’s Very Large Telescope in the Atacama Desert, Chile, and achieving 10 cm/s precision to detect earth-size planets. It is indeed amazing that current technology allows us to measure such tiny velocity changes with precision.

Although the very first exoplanet discoveries were based on RV measurements, the precision of the technique of those days was far less. In detecting HD 114762b, Latham and colleagues used a digital speedometer designed by the CfA specifically to directly measure stellar radial velocities, which had a precision of only about 1 km/s, two to three orders of magnitude less precise than what is possible today. In fact, the paper of Latham and company in Nature was only titled “The Unseen Companion of HD 114762—A Probable Brown Dwarf”, but with a note in the abstract that the companion “might even be a giant planet”. Latham has noted in his recounting of the discovery recently, “As long as there was only one radial velocity candidate, it was easy to dismiss it as unlikely to be planet.… It took seven years for this situation to change. The discovery of 51 Pegasi opened the floodgates for the discovery of a population of radial velocity planet candidates.”

Planet transit

The other widely used technique is the planet transit, which is when the planet passes between the (observer on the) earth and the host star. The planet blocks a small fraction of the starlight because of which there is a periodic dip in the star’s brightness (Figure 5) by a small amount of about 100 parts per million, which can last for one to 16 hours (an earth-like star will cause a dip of 84 ppm in the light from a sun-like star.) All transits of the same planet will produce the same amount of change in brightness and will last the same amount of time. It is thus a highly repeatable and robust detection method. This minuscule drop in brightness can be measured using precision photometry, the technique used to measure the amount of light emitted by stellar objects.

From the transit data, the following important planet characteristics are obtained: its size, from the brightness change and size of the host star; its orbital period, from the time between transits; and its orbital size, from the period and the mass of the star. The planet’s characteristic temperature can be calculated from the orbital size and the temperature of the star. The temperature is an indication of whether the planet is habitable or not. But the mass of the plant and its density, whether it is rocky or gaseous, can only be obtained when transit photometry is combined with RV measurements.

For a transit to be observable from the earth, the orbit must be aligned edgewise to the observer, which is not the case for the RV method. The probability of an orbit being properly aligned is equal to the diameter of the star divided by the diameter of the orbit. This is 0.5 per cent for a planet in an earth-like orbit about a sun-like star. Indeed, Kepler’s findings are all based on observing planetary transits across their host stars. The photometric precision of its charge-coupling device (CCD)-array-based instrument is 20 ppm. The minimum required for Kepler to detect a potential candidate are three or more transits of a given star, all with a consistent period, brightness change and duration.

Planet candidates

Discoveries by Kepler are called “planet candidates” or KOIs because they have not been verified as true planets. There are astrophysical configurations that can mimic the planet signal of the periodic dimming of the host star. One such configuration is the “eclipsing binary” star, in which the orbit plane of the two stars of the system lies so nearly in the line of sight of the observer that when the components undergo mutual eclipses it looks as if the starlight dips owing to planetary transit. Therefore, confirmation and verification are required through follow-up ground-based observations of Kepler data, using RV tracking of the potential candidate in particular.

The Kepler Mission has a detailed Follow-up Observing Programme of its findings using several ground-based telescopes. Independent verification by other ground-based programmes such as HARPS also helps to confirm planets from among the KOIs. “But, altogether,” Francois Fressin of the CfA was quoted in the Centre’s January 7 press release, “these [mimicking configurations] can account for [only] one-tenth of the huge number of KOIs.” This means 90 per cent of the KOIs should turn out to be bona fide planets, but confirmation being a long process, verification of such a large number will take years.

PRL, Ahmedabad

Besides making exoplanet discoveries entirely by themselves, the RV programmes around the world thus also have an important complementary role of verifying the current rapidly increasing number of exoplanet candidates from space-based observations.

The Physical Research Laboratory (PRL) in Ahmedabad too has launched a modest RV-based dedicated exoplanet detection programme, called PRL Advanced Radial-velocity All-sky Search (PARAS), using its 1.2-m infrared telescope at Mount Abu mounted with an indigenously designed and built precision spectrograph. In the coming years, this will be upgraded by using the instrument in conjunction with a larger 2.5-m aperture telescope that is proposed to be built (see box).