The MOA Project (Microlensing Observations in Astrophysics)
Ever wonder why coffee-table books on astronomy never show pictures of stars as sharply defined spheres like the Sun? After all, a moderate-sized telescope receives about a trillion photons (particles of light) per minute from a nearby star, easily enough to form a well-defined image.
Two phenomena prevent such an image from being obtained. First, if the telescope in question is ground-based, the photons that reach it are scattered slightly by turbulence in the Earth’s atmosphere, which blurs the image. This problem can be overcome by placing the telescope above the atmosphere. This was one of the prime motivations for building the Hubble Space Telescope, but even this ingenious creation falls foul of the second phenomenon.
According to the laws of quantum theory, for every photon recorded by a telescope there is a tiny but irreducible uncertainty in its arrival direction. Given Earth’s remoteness from the stars, this tiny angular uncertainty translates into a large spatial uncertainty back at the star of origin. This means that, at best, stars appear as hazy blobs, even when viewed with the Hubble, and that the light from any planets orbiting a star is lost in the glare of the hazy blob.
With these problems in mind, astronomers are developing special techniques for resolving stars and for detecting planets orbiting them (known as extrasolar planets), especially small, habitable planets like Earth.
The first detection of an extrasolar planet took place in 1995. It was achieved by two Swiss astronomers who observed perturbations to the motion of a star due to a giant Jupiter-sized planet orbiting nearby. Since then nearly 200 Jupiter-sized planets have been found in this way, as well as seven smaller planets comparable in size to Neptune. The presence of some of these planets has been confirmed by the dimming effect they cause when they pass in front of their parent stars. But such means of detection, especially when employed terrestrially, cannot reveal planets as small as Earth.
Enter the MOA project. This is a Japanese–New Zealand enterprise that uses the gravitational bending of light, as implied by Einstein’s General Theory of Relativity, to detect extrasolar planets. The technical term for this bending effect is gravitational microlensing, and the acronym MOA stands for Microlensing Observations in Astrophysics. Gravitational microlensing provides the most sensitive means presently devised for detecting extrasolar planets, and the only means of detecting planets as small as Earth orbiting normal stars like the Sun.
Earth-like extrasolar planets are, of course, the prime goal in the search for life beyond our solar system. For a planet to be habitable, it is generally believed it needs to be of sufficient size to retain an atmosphere; to have a rocky surface (not gaseous like Jupiter) to permit life to evolve on the surface; to be at a distance from its parent star that allows liquid water (not ice or steam) to be present on its surface; to have a parent star sufficiently long-lived that there has been time for life to evolve; to have a neighbouring giant planet to attract, and absorb the impact of, large asteroids that might otherwise hit it; and, possibly, to have a large moon to stabilise its spin axis. These, at any rate, are thought to be the likely requirements for life as we know it. Other types of life in different habitats might also be possible.
The study of extrasolar planets is not driven merely by the search for extra-terrestrial life, however. Indeed, one of the lessons learned since 1995 is that there is a wide diversity of types of planetary system. By observing the full range, astroscientists will understand better the physical processes involved in planetary formation and evolution.
That light was bent by gravitational fields was dramatically confirmed by the British astrophysicist Arthur Stanley Eddington and others in 1919. During a solar eclipse, stars close in the sky to the Sun, and therefore normally lost or hard to see in its glare, were clearly visible, and careful measurements showed their apparent positions changed in accordance with Einstein’s General Theory of Relativity. A contemporary re-evaluation of Eddington’s data reveals a less conclusive result, but more accurate measurements since then have put Einstein’s gravitation bending of light beyond doubt.
Einstein further considered the bending of light by stars other than the Sun in a short paper in 1936. He pointed out that if two stars, as viewed from the Earth, were lined up one behind the other, light from the more distant one would be bent by the gravitational field of the nearer one. As a result, the more distant star, rather than being hidden from view, or occulted, would appear as a ring of light surrounding the nearer star. Einstein referred to this effect as the “lens-like action of the gravitational field” of the nearer star, because the image, although magnified, would also be distorted. But incorrectly he predicted the effect would never be observed, because—and here he may have been influenced by the 1919 observations which could be made only when there was an eclipse, because of the Sun’s proximity to the Earth—the ring would be lost in the blinding glare of the nearer star.
In the 1960s, the idea was resurrected by Sjur Refsdal in Norway and Sydney Liebes in the US. Both men estimated that magnifications as high as 1000 could occur, and, in contrast to Einstein, held that distant faint stars and clusters of stars could act as gravitational lenses. In reality, background star, lens and Earth are rarely in the perfect alignment necessary to create a perfect Einstein ring, and because rays that pass closest to the lensing mass are bent more than rays passing further away, the ideal Einstein ring is generally reduced to two magnified blob-like images on opposite sides of the lensing star, as shown on page 81 (top panel). The two images are so close together that only their summed brightness is recorded by a telescope. As the lens star glides across the line of sight, the summed brightness displays a symmetrical increase and decrease. Planets orbiting the lens star act like smaller gravitational lenses to transitorily enhance or diminish the brightness of the observed star. Happily, the lensing zone is about the size of our solar system, and the probability for orbiting planets to cause an observable effect is quite high, especially if the lens star passes almost directly in front of the background star, and if intense observations are made during the approximately 24-hour period when this happens.
It is difficult to demonstrate gravitational microlensing directly because the bending is only slight and the Einstein ring is too small to be resolved. Artificial enlargement of the effect using plastic lenses milled to mimic the gravitational field can help. Both Liebes and Refsdal made such lenses, shaped like the bases of wine glasses, and the MOA group recently made two more. Simulated stars have been observed through these lenses under a variety of conditions.
Astronomers in New Zealand are ideally situated to observe gravitational microlensing. The densest star field in the sky—the centre of the Milky Way Galaxy, located in the constellation Sagittarius—passes almost directly overhead during the country’s long winter nights, delivering optimum conditions for microlensing. On any winter’s evening, several microlensing events are observable. A Polish–US group known as OGLE (the Optical Gravitational Lensing Experiment) made the first observation of microlensing in this region of the sky in 1993.
Other favoured sites for microlensing are the Large and Small Magellanic Clouds (two dwarf galaxies that orbit the Milky Way Galaxy). A US Australian group and a French group, known respectively as the MACHO (MAssive Compact Halo Objects) Project and EROS (Expérience pour la Recherche d’Objets Sombres), reported the first searches in these galaxies in 1993. They sought evidence of dark astrophysical bodies (rogue planets, dead stars, black holes, or brown dwarves—stellar bodies, with a mass lying below the thermonuclear ignition limit) in the Milky Way acting as gravitational lenses of stars in the more distant Magellanic Clouds. They were prompted to do so by the high speeds of stars in the outskirts of the Milky Way. These stars seemed to require the presence of a large amount of dark matter to confine them in the gravitational field of the Milky Way. After five years, however, the two groups reported that only a small fraction of dark matter could be in the form of MACHOs.
The Early searches for micro-lensing events by OGLE, the MACHO Project and EROS were conducted from observatories in Australia and Chile.
It was soon realised that New Zealand, too, was a location from which valuable observations could be made, serving as a hedge against cloudy weather at critical times in Australia and enjoying a large time differential with Chile.
At the time, Japan and New Zealand were collaborating on the observation of cosmic rays from a supernova that had been spotted in the Large Magellanic Cloud in 1987, and had established a good rapport for innovative research in astrophysics. In addition, New Zealand received an offer of support from the MACHO Project, which was led by Charles Alcock, formerly a student of the University of Auckland and today director of the Harvard-Smithsonian Centre for Astrophysics. John Hearnshaw, director of the University of Canterbury’s Mt John observatory, was quick to appreciate the scientific potential and offered observing time on one of the Mt John telescopes.
The rounded shape of Mt John makes for a smooth airflow over it and thus good astronomical “seeing”. To the west, the Southern Alps attract most of the precipitation that comes off the Tasman Sea, leaving Mt John relatively dry. Two of New Zealand’s foremost opticians, Norman Rumsey and the late Gary Nankivell, offered their expertise to widen the field of view of the telescope, and astronomers at Perth Observatory, Western Australia, offered to help computerise it.
Thus it came about that New Zealand and Japan commenced a gravitational-microlensing project at Mt John. The name “Microlensing Observations in Astrophysics” was proposed by Japan, and New Zealand readily accepted it. More importantly, the main aim of the project was agreed upon: to search for dark matter and extrasolar planets as revealed by gravitational-microlensing. Observations began in 1995.
The years 1995 to 2000 were a learning period for the MOA project, as they were for others in the same business. In retrospect, it will probably become clear that we are still in a learning phase. One of the first skills we had to develop was the analysis of images containing millions of overlapping stars. Techniques were also required for modelling complex microlensing events using clusters of computers. And observations were made every fine night, about 60 per cent of night-time hours.
Perhaps the main lesson learned by MOA in the early years was the advantage of observing microlensing events involving high magnifications, of the order of 100 or more. Rutherford learned a similar lesson almost a hundred years ago. In his classic experiment of 1910, he fired alpha particles (ionised helium atoms) at a sheet of gold foil. Most of the particles, as expected, passed straight through, but some were deflected, having struck something very small and dense at the heart of atoms in the foil. The effect was comparable to gravitational microlensing. The obvious difference is that in Rutherford’s experiment particles were deflected away from the centre of an atom, while microlensing entails photons being attracted by gravity. However, inserting a simple minus sign into the mathematics takes care of this. The important point is that in both situations the most useful information is obtained when the projectile originates directly behind the obstacle. In Rutherford’s case it led to his discovery of the atomic nucleus. For the MOA project and others of its kind, such an alignment yields the largest magnifications in gravitational microlensing.
The gravitational-microlensing event MOA-2002-BLG-33 provides an illustration of what can be achieved when magnification is high. As its name records, this was the 33rd microlensing event observed by the MOA group in the central bulge of the Milky Way during 2002. It reached a peak magnification in excess of 500. The lens for the event was a binary star; that is, a pair of stars orbiting each other about a common centre of mass. Such stars are common, and they form beautiful lenses.
An illustration on the preceding page shows the kite-shaped magnification pattern produced by the binary lens of MOA-2002-BLG-33, and also the star that was magnified. This magnification pattern is typical of the complicated bending of light that occurs with lenses containing more than one component. The star passed horizontally through the kite, taking 15.6 hours. Intense observations made during these hours enabled the characteristics of the star to be measured without the high degree of uncertainty explained above, as if Sherlock Holmes had passed a large magnifying glass over it. By making measurements through different coloured filters, it was possible to form a colour image. The angular resolution of this image is probably the finest yet achieved in any situation and by any means. It corresponds to the resolution needed for this article to be legible from Earth if it was placed on the Moon. The MOA group was assisted in this work by observatories in Chile, South Africa, Israel and Arizona, as well as by the Hubble Space Telescope.
MOA-2003-BLG-53 was another highlight for the MOA project. This was the first microlensing event in which irrefutable evidence of an extra-solar planet was obtained. The planet found in this event is quite similar to Jupiter. It’s mass is double that of Jupiter, and it orbits a star two thirds the mass of the sun in an orbit slightly larger than that of Jupiters’. This star-planet system is some 19,000 light years away, about three quarters the distance to the centre of the Milky Way. The planet was found by the MOA group and by the OGLE group observing from Chile. The combined data from Mt John and Chile are shown overleaf, plus an image of the event obtained by the Hubble Space Telescope in 2006. This shows that in three years since microlensing occurred, the lens and background stars diverged sufficiently for the Hubble to just resolve them, through their different colours. In the future, similar follow-up observations will be made routinely as they enable host stars of planets found by micro-lensing to be classified.
An event found by OGLE in 2005 provided a good example of the power of gravitational microlensing at high magnification. This was OGLE-2005 BLG-169, which reached a magnification in excess of 700. Measurements of the intensity of the magnified star as its light passed through the lens showed that the lens consisted of a small cool star, or red dwarf, that was orbited by a planet, comparable in size to Neptune, at a distance about four times that of the Earth from the Sun. The estimated temperature of the planet was about –200°C. It was possible to determine that it had no gas-giant companions similar to Jupiter, suggesting that it was itself free of gas. The painting on the opening spread of this article is an artist’s rendering of this planet.
In 2002, Yasushi Muraki, professor of physics at Nagoya University and one of New Zealand’s long-standing collaborators from Japan, was awarded a Japanese government research grant to build a new, larger telescope at the Mt John observatory for the MOA project. This was welcome news, as the telescope the MOA project had been using since its inception was significantly smaller than those used by other microlensing groups. The telescope was funded for the purpose of searching for both dark matter and planets in the Milky Way.
Andrew Rakich, a protégé of Nor‑man Rumsey working at the time for Industrial Research Ltd in Wellington, was selected to design the telescope, and Nishimura Optical Company of Kyoto was commissioned to build it. Eco-tourism and education company Earth & Sky, of Lake Tekapo, offered to supply the control room. A design promising 15 times the light-gathering power of the telescope used to date was agreed on, and construction began in 2003. The telescope was fully assembled in Kyoto, then disassembled, shipped to New Zealand, and reassembled in 2004.
While the telescope was under construction, a large electronic camera incorporating a charge-coupled device, or CCD, was designed and built at the University of Nagoya under the direction of Takashi Sako. The camera is the heart of the telescope—all data are recorded by it. Modern CCD cameras have transformed astronomy through their high (almost 100 per cent) light-detection efficiency and their compatibility with powerful computers. They enable millions of stars to be monitored simultaneously. High accuracy is achieved by operating at low temperatures, around –90°C, in a vacuum. The construction of these large, specialist cameras is as complicated as the construction of the telescopes themselves.
An opening ceremony for the new MOA telescope was held on December 1, 2004. Among the 150 guests and dignitaries who attended were Masaki Saito, Japan’s ambassador to New Zealand, Professor Shinichi Hirano, the president of Nagoya University, and Yuji Nishimura, CEO of Nishimura Optical Company. Whetu Tirikatene-Sullivan, wife of MOA team member Denis Sullivan, blessed the telescope and thanked the Japanese government for providing the funding for it.
In 2005 exhaustive commissioning tests were carried out, culminating in the somewhat serendipitous discovery, in association with other micro-lensing groups, of the most Earth-like extrasolar planet yet found. This was an icy planet, similar to OGLE-2005 BLG-169, but slightly smaller.
The future appears bright for the MOA project, and indeed for the microlensing community in general. The ability to detect extrasolar planets as small as Earth means that the abundance of such planets can be measured. From there it should not be a difficult step to estimating the abundance of habitable planets in the universe—an exciting prospect.
To achieve this goal, the MOA project and OGLE will continue their surveys of the Galactic bulge and the Magellanic Clouds for microlensing events. The most promising of these, in particular those with the highest magnifications, will be selected for especially close scrutiny by networks of “follow-up” telescopes.
Two such networks have already been in operation for some years: PLANET/Robonet (PLANET = Probing Lensing Anomalies NET-work) and MicroFUN (Microlensing Follow-Up Network). They have telescopes in Australia, South Africa, Israel, Hawaii, the US, Chile and New Zealand. The Neptune-like extrasolar planets mentioned above were found in collaboration with these groups and OGLE. Of special note here are important observations made of the peaks of high-magnification events from the Stardome and Farm Cove observatories in Auckland using the smallest telescopes in the microlensing community.
More recently, astronomers from Taiwan, China, Japan, India, Thailand, Namibia, Cuba, Venezuela and Argentina have expressed interest in observing future events. Gravitational microlensing is in a growth phase!
Looking further to the future, Antarctica offers exciting prospects. Australian and French astronomers have recently found that the atmospheric conditions at high points on the Antarctic plateau are the best for astronomy on the planet. Images of Scott battling blizzards do not reflect typical plateau weather. Dry, stable conditions are the norm—every astronomer’s dream. The scintillation, or twinkling, of stars is less by a factor of three than at the best sites in Chile and Hawaii. A large French–Italian station has been operating successfully at one of the high points, Dome C, since 2005, and last year a Chinese group reached the highest point on the plateau, Dome A. A telescope at either of these locations would be ideally placed to monitor gravitational microlensing events continuously.
Another possibility is to add further telescopes to the battery that already exists at mid-southern latitudes. Andrew Gould of the US, one of the pioneers of the observational era of gravitational microlensing, has proposed the construction of two further telescopes similar to the MOA one, to be located in South Africa and Chile. And David Bennett, another pioneer of observational microlensing from the US, has proposed the construction of a space telescope dedicated to gravitational microlensing. This could be used to make an accurate census of all types of planet in the Milky Way Galaxy.
The hunt for Earth-like planets is spurred both by scientific curiosity and by competition among various groups using different techniques. Gravitational microlensing is not the only game in town. A US space mission named Kepler will shortly commence a search for Earth-like planets orbiting nearby stars using the “transit” technique, which involves detecting where light from a star is dimmed by planets orbiting it. Kepler will be most sensitive to hot planets orbiting quite close to their parent stars. Gravitational microlensing, on the other hand, is most sensitive to cold planets orbiting further out. The combined results will provide valuable crosschecks, and yield unbiased information on warm Earth-sized planets in the so-called habitable zone.