- Email: [email protected]
The Darwin mission: Search for extra-solar planets
Advances in Space Research 36 (2005) 1114–1122 www.elsevier.com/locate/asr
The Darwin mission: Search for extra-solar planets L. Kaltenegger a
a,*
, M. Fridlund
b
Science Payload and Advanced Concepts Office, European Space Agency ESTEC, P.O. Box 299, NL-2200AG Noordwijk, The Netherlands Research and Scientific Support Department, European Space Agency ESTEC, P.O. Box 299, NL-2200AG Noordwijk, The Netherlands
b
Received 14 January 2005; received in revised form 19 May 2005; accepted 19 May 2005
Abstract The direct detection of an Earth-like planet close to its parent star is challenging because the signal detected from the parent star is between 109 and 106 times brighter than the signal of a planet in the visual and IR respectively. Future space based missions like Darwin and TPF-I concentrate on the mid-IR region between 6 and 20 lm, a region that contains the CO2, H2O, O3 spectral features of biomarkers in EarthÕs atmosphere. The InfraRed Space Interferometer Darwin is an integral part of ESAs Cosmic Vision 2020 plan, intended for a launch towards the middle of the next decade. It has been the subject of a feasibility study and is now undergoing technological development. It is focused on the search for, and characterization of Earth-like planets orbiting other stars. A secondary objective is to carry out imaging of astrophysical objects with unprecedented spatial resolution. The implementation is based on the new technique of Ônulling interferometryÕ. New designs have been developed that will be implemented on four spacecrafts and search for planets around a minimum of 165 stars within the mission lifetime. 2005 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Darwin; Extrasolar planet search; Biomarkers; Nulling interferometry; Darwin targets stars; TPF-I
1. Introduction The InfraRed Space Interferometer Darwin is a major element in the Cosmic Vision 2020 program of the European Space Agency. Darwin has the explicit purpose of detecting other Earth-like worlds, analyze their characteristics, determine the composition of their atmospheres and investigate their capability to sustain life as we know it. The closing years of the 20th century have allowed us, for the first time, to seriously discuss interferometric instruments deployed in space achieving unprecedented spatial resolution. These missions will lead to new astrophysics. Especially – and this is the greatest challenge – * Corresponding author. Present address: CfA, Astronomy Department, 60 Garden Street, Cambridge, MA 02138, United States. Tel.: +1 617 495 7158. E-mail address: [email protected] (L. Kaltenegger).
we expect to be able to carry out the first detailed study of terrestrial exoplanets (defined as planets similar to our own Earth as what concerns size and mass, and orbiting other stars than our Sun) as well as comparative planetology based on the diversity of planets detected. The detection and study of the latter promises to open a new era in science and will affect a broad spectrum of science and technology. We can now confidently expect the first results from space based interferometers within 10 years. Sophisticated instruments will follow in short order. Analysis of the planetary light requires that the stellar light is suppressed to a high degree. This is done by a technique called nulling interferometry, in essence this means that achromatic phase shifts are applied to the beams collected by individual telescopes before recombination such that the on-axis light, i.e., stellar light, is cancelled by destructive interference, while the much weaker planetary light emitted at a certain off axis angle interferes constructively.
0273-1177/$30 2005 COSPAR. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.asr.2005.05.061
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
The Darwin mission is implemented on four free flying spacecraft including one beam-combining spacecraft. The beam combiner and the telescope spacecraft fly in one plane with each telescope spacecraft at the same distance from the beam combiner. The resolution of the interferometer is adjusted by changing the distance between the telescope spacecrafts. A similar activity has been taking place in the United States within the context of NASAs Origins program. A science collaboration has already been established.
2. Extrasolar planets The last years have seen the detection of planets beyond our solar system finally becoming a fait accompli. The techniques utilized so far have been indirect. Radial velocity relies on measuring the reflex motion of the parent star, with respect to the common center of mass of the star–planet system. The transit method measures the reduction in flux when a planet crosses the line of sight between earth and its host star. The first detection of a planet outside the Solar System using radial velocity method was reported by Mayor and Queloz (1995) for the Solar type star 51 Peg. This was quickly followed by the Lick group who reported planets around 70 Vir and 47 Uma (Marcy and Butler, 1996; Butler and Marcy, 1996). It appears unlikely, however, that this method could be used to infer the presence of Earth-type planets. The current precision in radial velocity is 3 m s 1, while the Earth at 1 AU from the Sun would require 0.1 m s 1. Acoustic pressure-mode oscillations in solar-type stars have amplitudes of 0.5–1 m s 1, and will thus make it essentially impossible to detect a deflection of an Earth simile. The planets so far found with the radial velocity method are objects more akin to the planet Jupiter than something like our own Earth. Radial velocity searches for planets are strongly biased toward planets with large mass and short orbital period because of their easy detectability. As time passes longer and longer periods are being picked up, and there are now a number of confirmed planets in orbits with periods of several years. A few of the objects are likely Brown Dwarfs. The radial velocity method provides us with a lower limit to the mass and the orbital radius. If the inclination of the planetary orbit is not known, there is no way to determine the planets absolute mass, actual size or composition – without other data. One trend that has been seen in radial velocity star surveys is that a plot of the number of detected objects vs. M sin(i) increases strongly towards M sin(i) < 1 Mjup. An unknown process would have to bias our detection towards only picking up systems seen face on (Fischer and Valenti, 2003; Marcy et al., 2003). In the case of the first observed occultation, the planet orbits the star HD 209458 every 3.52 days at a distance of
1115
about 0.05 AU. The occultation lasts about 2.5 h and from this observation, the inclination is found to be 87.1. Charbonneau et al. (2000) derive a planetary mass of 0.63 MJup and a planetary radius of 1.27 RJup. This can be done since the orbital radius is well known from the radial velocity measurements, and the stellar radius is known to good accuracy from stellar evolution theory. The actual shape of the light curve during the occultation (Mazeh et al., 2000), constrains the planetary radius and orbital inclination to a very high degree. The average density of the planet turns out to be only half that of the major gaseous giant planets in our own Solar System immediately ruling out that the planet is a rocky, terrestrial body (super-Earth) that could have formed close to its current location since such a planet would be significantly smaller than 1.27 RJup. The planet is thus a gas giant. Being physically larger than Jupiter but with a lower mass is caused by its proximity to its primary which heats it to a surface temperature of 1200 K. Such temperatures would, however, only affect the outer 1% of the planet, and the large diameter immediately says something about its evolution. As pointed out by Lunine (2001), what is happening is that the flux from the star retards the cooling of the planetary interior. A giant planet formed in isolation would cool in a brief time (106 years), and thus it also shrinks rapidly from its original distended state. For a planet in very close proximity to a star such as is the case for HD 209458b, the atmospheric temperature profile is flattened and the rate by which heat can be transported outwards from the interior is reduced and the contraction will be retarded. Detailed models (Burrows et al., 2000) show excellent agreement with the planetary radius at its current age of 4–7 billion years (the age being determined from stellar evolution theory). It can also be shown in these models, that the planet must have arrived at an orbital radius of 0.05 AU within at most a few tens of millions of years after formation. Otherwise it would take longer than the present age of the Universe for the external heat to diffuse inwards far enough to expand the radius to the observed value. The observation of a single occultation thus shows that the so called Ôhot JupitersÕ either form in place or migrate inwards within at most a few · 107 years (Lunine, 2001; Burrows et al., 2000). Another indirect method is to obtain astrometric data and thus track a starÕs path across the sky, measuring the wobble introduced by the rotation around the common center of mass of the star–planet system. The European Space AgencyÕs GAIA mission promises large statistical surveys of massive planets (Perryman, 2000). 2.1. Direct detection of terrestrial planets All of the above mentioned methods will continue to refine our knowledge about planetary systems. Unfortunately they are restricted towards the indirect detection
1116
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
of relatively massive planets. Space missions (e.g., COROT, Eddington, Kepler) designed to study occultation will eventually pick up lower (Earth-size) mass bodies, but relatively little information will be gathered in this fashion. The most important datum will be the frequency of Earth-like planets. The main problem in the direct detection of an exoplanet of a size comparable to our own Earth – and located at a similar distance from its own star – is one involving contrast and dynamical range. A central star like our Sun (G2V) outshines an Earth-like planet in the visual wavelength range by a factor of at least 109. Going to the mid-infrared alleviates this problem, because the planets thermal emission peaks (a terrestrial planet is in this case defined as one of roughly the same size as the Earth and of the same surface temperature – ca. 270 K – thus having its peak emission at 10 lm). Even at these wavelengths, the contrast is more than a factor of 106. The star and planet will be very near each other on the sky, and we need to devise a way of extinguishing the light from the star. Different coronographic methods (in space) have been evaluated, and albeit having the capability of achieving the scientific objective of detecting the exoplanet, these methods do not lend themselves to a large enough search space (unless the telescope is extremely large). Lately it has been suggested to fly coronographic systems operating in the visual wavelength range in space. A 5–10 m diameter monolithic telescope could then suffice but would investigate a smaller search space than the interferometer proposed.
3. Nulling interferometry Infrared nulling interferometry seems to be the preferred instrumental concept for the implementation of a mission observing a large sample of at least 165 stars, for which technological readiness can be achieved in a time frame compatible with a launch during the second half of the next decade. A nulling interferometer extinguishes light from an on-axis source. The basic concept is to sample the incoming wavefront from the star and its planet with several telescopes that individually do not resolve the system. Nulling interferometry differs from ÔnormalÕ imaging interferometry (Michelson interferometry) in that one attempts to obtain a dark fringe (or fringe pattern) in the center of the field, by introducing phase shifts in the light paths of one or several of the interferometer arms. By keeping a star in the center of the image plane, coronography is realized without the presence of a physical mask. In its simplest form, a two-telescope nulling interferometer introduces a 180 phase-shift between the two apertures, resulting in destructive interference along the line of sight. At the same time, light at
small off-axis angles from the line of sight will experience constructive interference, thus allowing a faint object close to a bright star to be discernible. The observable area around the central null depends on the separation of the telescopes – the baseline. The star can be centered on the deep central null. Specific off-axis locations like the position of an earth-like planet orbiting the host star can be constructively interfered by adjusting the distances between telescopes. The output of the system can be described by an angular transmission map (TM) featuring interference fringes, with a sharp null (destructive interfered area) in the center of the map. The stellar signal is nulled out only on the optical axis. A leakage of photons out of the central null exists because the star has a finite photospheric disk. That leakage is a very important noise source. Information about the distribution of planets in the target system can be recovered by modulation of that signal. No spatial information is extracted in a single exposure. Rotation modulates the interferometer output intensity as a planet passes in and out of the dark fringes. From the intensity and actual pattern of this modulation one can derive the planetÕs parameters see e.g., Ollivier et al. (2003) and Kaltenegger and Fridlund (Kaltenegger and Fridlund, 2004). The use of more telescopes achieves a symmetric pattern around the star, with a deep central null placed on the starÕs position. The actual shape and transmission properties of the pattern are a function of the number of telescopes, configuration, and the distance between the telescopes (Mennesson and Mariotti, 1997; Absil, 2001; Kaltenegger, 2004). In practice, the central null of the transmission map will be degraded by amplitude and optical path differences between the interferometer arms what will lead to a level of noise that sets a limit for the achievable null depth. The properties of the Bracewell array in the presence of phase and amplitude perturbations have been discussed by Serabyn (2000). The use of a single Bracewell as nulling interferometer for planet detection is limited by the impossibility to rapidly modulate the planet signal against disturbances such as stellar leakage and the background. For a system at 10 pc observed at 12 lm the local zodiacal cloud signal is the biggest noise source, followed by the stellar leakage and the signal of the exo-zodiacal dust disk. Here, we assume an exozodiacal dust disk similar to our own system. In order to separate out the signal of the planet from this background one needs to modulate the planetÕs signal and that from the zodiacal dust at different frequencies. The planetary signal is separated out from that of any extrasolar zodiacal dust because of its temporal and spectroscopic behavior (Bracewell and McPhie, 1979; Leger et al., 1996; Mennesson and Mariotti, 1997; Absil, 2001) due to modulation. The modulation technique, envisioned is internal modulation. It is based on
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
combining the outputs of various sub-arrays under a variable phase shift (see e.g., Absil et al., 2003). Alternative modulation techniques are under investigation (dÕArcio et al., 2004) especially in the context of GENIE, a proposed ground-based pre-cursor for DARWIN at the VLTI. The influence of exo-zodiacal clouds is an important constraint for planet detection with DARWIN. This issue will be addressed by Ground-based European Nulling Interferometry Experiment (GENIE) (Gondoin et al., 2004). The prime objective of GENIE is to gain experience with the design, construction and operation of a nulling interferometer, as preparation for the DARWIN mission. GENIE will be particularly sensitive to warm circumstellar dust as it operates at mid-infrared wavelengths and thus provides the opportunity to investigate the properties of the DARWIN target stars (Absil et al., 2003; Kaltenegger and Karlsson, 2004). Fig. 1 shows that the detected signal is governed by the contribution of background noise to the interferometer output signal (where star leakage and EZ referee to the residual signal from the star and the exo-zodiacal dust disk respectively, and LZ and thermal referee to the flux from our local zodiacal cloud and the telescopes at 30, respectively). Fig. 1 shows the results for the Three Telescope Nuller (TTN) architecture (Karlsson et al., 2004) (see explanation in Section 4). The output beam comprises all information from a specific set of sources on the sky (consisting of star, planet(s) and zodiacal dust surrounding the target star), as well as the background. The local zodiacal cloud (LZ) provides the foreground through which DARWIN will observe. Because zodiacal background is diffuse, it cannot be cancelled by nulling interferometry. The stellar leakage is dominating for short wavelengths, while the thermal background level due to the emission of the optics, can approach that of the zodiacal dust for temperatures above 40 K at long wavelengths. However, the LZ dust disk remains the biggest noise factor, for most of the wavelength band if we assume that the exo-zodiacal
Fig. 1. Background and earth-sized planet for a G star system at 15 pc as a function of wavelength (Kaltenegger, 2004). The graphs show the modulated signal as obtained for a triangular TTN configuration (Karlsson et al., 2004), including detector characteristics and fiber coupling effects. The planet is shown on a transmission maximum.
1117
cloud (EZ) is similar to our own LZ cloud. Note that large coherent structures such as wakes and clumps behind planets can masquerade as planets. Those structures could also serve as markers for the presence of planets, if their location in respect to a planet were well understood. As exo-zodical clouds are not uniform, a planet must be detected against a non-flat field of corrugations. Structures in our own cloud have roughly <0.1% of the amplitude of the total cloud brightness (Backman et al., 1998) and are thus no source of confusion. At levels less than 1% of the total brightness in a simulated DARWIN observation scenario, the structured emission is not a significant contributor to the noise budget. In an ideal scenario with only a single planet around its host star and no other disturbing sources, such as extrasolar zodiacal dust in the target system, the detection of a positive flux would imply that a planet is present, if the star is well and truly Ônulled outÕ. In real observations, several factors affect the signal to noise in a detrimental way. This has been analyzed by different groups e.g., Velusamy et al. (2003) and Kaltenegger and Karlsson (2004).
4. Architecture design The original DARWIN mission concept was optimized for stellar rejection to focus on observing the closest stars for planetary companions. This led to a baseline concept of a free flying configuration with six collector telescopes of 1.5 m diameter and a central beam combiner. A number of alternative mission architectures have been evaluated on the basis of interferometer response as a function of wavelength, achievable modulation efficiency, number of telescopes and starlight rejection capabilities (Kaltenegger, 2004; den Hartog et al., 2004; Dubovitsky and Lay, 2004). These analyses show that the starlight rejection criteria can be relaxed while still maintaining a target sample of 165 stars. The integration time for close by stars increases, but the overall mission performance does not degrade significantly. The leakage term and the resulting null degradation depend strongly on the baseline used. Optimizing the baseline of the interferometer for each stellar target system will minimize the leakage, while a fixed baseline as would be used for an interferometer implemented on a structure would lead to a high level of leakage for most of the target systems (distances vary between 3 and 25 pc) and have a variation of optimized baselines. Accordingly, candidate Darwin and TPF configurations use three (Karlsson et al., 2004) or four telescopes of 3.5 m diameter, reducing complexity and cost of the mission. Three telescopes is the minimum number of telescopes needed for an interferometer mission that
1118
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
uses rapid signal modulation to detect a planet in the high background noise. The baseline mission duration is 5 years, extendable to 10 years in an L2 orbit. Operating in the infrared band requires that all optical components are cooled to roughly 40 K, this is achieved by passive cooling. Only the detector requires active cooling. The first practical demonstration of nulling was undertaken in February 1998 (Hinz et al., 1998). Using the Multiple Mirror telescope on Mount Hopkins, Arizona, Angel and his team were able to cancel out the image of a star: a-Orionis. The nulled image had a peak in intensity of 4.0% and the total integrated flux of 6.0% of the constructive image.
5. Biomarkers The spectral range between 6 and 18 lm and preferably 5–20 lm has been selected for DarwinÕs extra-terrestrial planetary research since the ratio between planetary and stellar light is improved by roughly three orders of magnitude compared to the visible spectrum and since absorption bands of important tracers of life (H2O, O3 and CO2) are present in this range (Ollivier these proceedings). A major goal of the mission is not only to detect terrestrial exoplanets, but also to investigate if the conditions on the planet in question would allow life as we know it to exist and indeed if it already exists. In order to do so we need to define what life is, and how life as we understand it interacts with its environment in an observable way. In the mid-IR the classical signatures of biological activity on Earth are the 9.6 lm O3 band, the 15 lm CO2 band, and either the 6.3 lm H2O band or its rotational band that extends from 12 lm out into the microwave region or both. In the same spectral region, the 7.7 lm band of CH4 is a potential biomarker for early-Earth type planets. The 9.6 lm O3 band is highly saturated and is thus a poor quantitative indicator, but an excellent qualitative indicator for the existence of even traces of O2. Observations from 8 to 12 lm of the H2O continuum allow estimates of the surface temperature of Earth-like planets. However, one should note that the atmosphere of planets that are warmer than about 310 K will be opaque in this region because of continuum absorption by water vapor. While the existence of H2O in liquid state on the surface of a planet is considered essential for the development of life, it is not by itself a bio-indicator. The presence of CO2 in the atmosphere of all terrestrial planets in our Solar System is of course inconclusive as what concerns signs of biological activity but provides information about the physical state of the atmosphere. Taken together with molecular oxygen, abundant CH4 can indicate biological processes, although depending
on the degree of oxidation of a planetÕs crust and upper mantel also non-biological origins can under certain circumstances produce large amounts of CH4. With a spectral resolution, Dk = 20–40, we can detect all of the features mentioned above. We have to remember, that this refers to our Earth as it is today. The oxygen level in our atmosphere has been relatively high only for the last 20–30% of the EarthÕs history. Although, it is now more or less generally accepted that life arose on the Earth immediately after the era of bombardment, i.e., 3.8 · 109 years ago, it remained in the sea, and at a relatively simple level until just about 600 million years ago. Life before the rise of oxygen in the atmosphere around 2.5 Gyr ago was dominated by Methane producing species and if we want to define our Ôremote sensingÕ criterion such that life is indicated by a disturbance of the equilibrium of a terrestrial planets atmosphere, we need to also take this evolutionary aspect into account. In order to interpret spectroscopic data in the context of possible biological activity we also need to know the planetÕs size and temperature. In the thermal part of the spectrum, the shape gives a measure of the temperature of the object examined. The mid-IR spectra can give the planetÕs albedo, the temperature of the observable emitting regions and thus the planetÕs size. If all life on the Earth were removed suddenly – bacteria, green plants, etc., all of the free oxygen in the EarthÕs atmosphere would disappear in the geologically short time of 4 million years. The atmosphere of the Earth is out of equilibrium as is evidenced by a comparison with models or with the situation in the other terrestrial planets in the Solar system – Mars and Venus. This disequilibrium is caused by the living things on our planet. Previous to life being dominated by Oxygen generating species, the atmosphere of the early Earth was out of equilibrium by Methane, CH4.
6. Life and habitability The search for life requires a general definition of life. Life contains information; life is self-replicating; life evolves; and life influences its environment. We are attempting to detect these attributes of life, through remote sensing, at interstellar distances. However, no clearcut observables can be indicated – particularly such that can be detected over interstellar distances. One should, however, remember, that based on the two first criteria, we would also determine that a computer virus is alive. We have to expand our criteria. Faced with this problem, space agencies currently base their development of missions on what life on our own planet looks like. To define search criteria, we presume that life has a carbon-based structure not too dissimilar from our own. Owen (1980) has argued that carbon chemistry with water as a solvent is the most
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
likely base of life. Carbon is a highly abundant element and exhibits a remarkable ability to form a host of highly complex molecules. Arguments in favor of water as a solvent are that it is highly abundant, an excellent solvent and remains liquid over a broad temperature range. The temperature range is high enough for chemical reactions to occur rapidly but also allows large molecules to form. Specific conformation of macromolecules can form when in solution in water due to the solventÕs attraction of hydrophilic groups and repulsion of their hydrophobic ones. These conformations allow very specific chemical reactions to build reproducible complex structures. Water dissociates to form oxygen and ozone which can protect liquid water from further dissociation, provided there is a cold trap in the atmosphere on the hypothetic planet to confine water to lower altitudes. Basing the search for life on the carbon chemistry assumption it is then possible to establish criteria for habitable planets in terms of their size and distance from their stars. Gases in their atmosphere such as oxygen, methane and water vapor would indicate in combination the presence of life (see Fig. 2). Based on the history of our own planet, atmospheric features can provide clues of possible life forms for at least 2 billion years. This is more than 107 times longer than radio signals reveal the presence of an advanced civilization on our planet. The criterion on temperature defines a habitable zone HZ (Kasting et al., 1993) around each individual star. The HZ around different stellar types will vary as it only depends on the spectral energy distribution of the star. The HZ around F stars is larger and occurs farther out, the HZ for K and M stars is smaller and situated farther in than around our Sun. High ultraviolet fluxes could be a problem for life around F stars. For stars later than about K5 spectral class the HZ falls within the 4.6 Gyr tidal locking radius of the star. The planetary companions may rotate synchronously with their orbit, resulting in a freezing out of the atmosphere on the dark side of the planet. The temperature difference between the bright and the dark side of the planet could generate winds that efficiently reduce the temperature gradient on the planet if the atmosphere were thick enough. A thick CO2 atmosphere would very likely cover bio-signatures like O3, what makes M stars secondary target because biomarkers that indicate the existence of life as we know it would not be detected. Strictly speaking, the surface temperature of a planet will depend not only on the energy input from the primary, but also on the atmospheric pressure and composition. Our own Earth, for instance, would be significantly colder without its green house effect, caused by CO2 and CH4. Since we a priori have no idea about the presence and/or composition of any eventual atmospheres, we will have to use this criterion with some care in an individual case (Selsis, 2004).
1119
Fig. 2. Thermal infrared spectra of Venus, Earth and Mars. The 15 lm CO2 band is seen in all three planets. Earth also shows evidence of O3 and H2O (ESA, http://sci.esa.int/science-e-media/img/c8/planetbig.jpg, 01.10.2004).
A planet with O3 and H2O and CO2 absorption bands in its spectrum that lies within the HZ is the main target of a mission that searches for habitable planets. We point out the possibility that life exists on planets that do not show H2O and O3, if the production of O2 by photosynthesis is not able to overcome the oxygen sinks, e.g., Mars-like planets have small O2-sinks and Venuslike planets large abiotic O2 sources (Kasting et al., 1993).
1120
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
7. Target star catalogue An early target list of stars, mainly from the Gliese and Jahreiss catalogue of nearby stars (1991) was compiled by A. Leger and M. Ollivier for the Darwin feasibility study for F, G and K stars. In preparation of the Darwin mission we (Kaltenegger, 2004) established an updated catalogue including new information and an extended star sample of stellar systems of nearby FGKM stars that are potentially habitable to complex life. The list was created from the Hipparcos catalogue by examining the information on distance, stellar variability, multiplicity, location and spectral classification as well as the interferometerÕs limitation. Four selection criteria were adopted: (i) FGKM stars within a distance of 25 pc; (ii) main sequence stars nat-
ure based on the B-V index; (iii) stars within a cone of aperture ±45 of the ecliptic plane; (iv) Stars with magnitudes smaller than 12 in V. Limitations imposed by the instrument are e.g., the 45 observation cone in anti-sun direction, a minimum distance to a close binary companion that cannot be simultaneously nulled while searching for a planet and a minimum observation time of the star in the field of view of the instrument. The 45 cone in the selection criteria is given by constraints of the sunshields on each free flyer. The shields will be designed to permit a ±45 cone of observation in the anti-sun direction. That also influences the maximum time during which a target can be observed. The selection criteria for the updated DARWIN stars from the HIPPARCOS catalogue lead to a total of 807 stars. The current analysis results in a prime target list of
Fig. 3. Spatial distribution of all prime DARWIN target stars (FGKM) (Kaltenegger, 2004).
Fig. 4. Spatial distribution of FGK prime DARWIN target stars (Kaltenegger, 2004).
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
1121
Darwin has the explicit purpose of detecting other Earth-like worlds, analyze their characteristics, determine the composition of their atmospheres and – investigate their capability to sustain life as we know it, an exciting challenge that we are prepared to answer during the second half of the next decade.
References
Fig. 5. Distribution of DARWIN prime targets over distance (Kaltenegger, 2004).
285 target stars of stellar type F, G, K and selected M stars excluding multiple systems and active stars, consistent with the selection criteria within 25 pc of our sun (see Figs. 3–5). For the G type stars the properties of the stellar targets were revised (Eiroa et al., 2003) through consultation of existing data archives. Note that the HIPPARCOS catalogue is not complete to the 12th magnitude. As Darwin concentrates on brighter/closer star the completeness of the HIPPARCOS sample is an issue for densely populated areas. Thus, the number of stars quoted is a first analysis. Further data on nearby stars will be gathered and the target list will be updated once new data is available. The free flyer array architecture can be optimized to detect planets in the Habitable zone around the selected target stars. This is a big advantage of a free flying array over a structurally connected interferometer.
8. Conclusion The Darwin model mission represents today a design that could answer one of the longest lasting scientific questions – if we are alone in the Universe. The feasibility study has demonstrated the maturity of the existing technology (Karlsson and Kaltenegger, 2003). The design assessment leads to a design consisting of four free flying spacecrafts that can observe a minimum of 165 stars from the DARWIN target star list (Kaltenegger, 2004). Infrared nulling interferometry emerges as the preferred instrumental concept for the implementation of a mission observing a large sample of at least 165 stars, for which technological readiness can be achieved in a time frame compatible with a launch during the second half of the next decade.
Absil, O. Nulling Interferometry with IRSI Darwin: Detection and Characterization of Earth-like Exoplanets, Master Thesis, Universite´ de Lie`ge, 2001. Absil, O., Karlsson, A., Kaltenegger, L. Inherent modulation: a fast chopping method for nulling interferometry. Proc. SPIE 4852, 431– 442, 2003. Backman, D.E., Caroff, L.J., Sanford, S.A. et al. (Eds.), Proceedings of the Exozodiacal Dust Workshop (NASA/CP-1998-10155), 1998. Bracewell, R.N., McPhie, R.H. Searching for nonsolar planets. Icarus 38, 136–147, 1979. Butler, R.P., Marcy, G.W. A planet orbiting 47 Ursae Majoris. ApJ 464, L153, 1996. Burrows, A., Guillot, T., Hubbard, W.B., Marley, M., et al. On the radii of close-in giant planets. ApJ 534, L97, 2000. Charbonneau, D., Brown, T.M., Latham, D.W., Mayor, M. Detection of planetary transits across a Sun-like star. ApJ 529, L45, 2000. dÕArcio, L., Karlsson, A., Godoin, P. Use of OPD modulation techniques in nulling interferometry. SPIE 5491, 2004. den Hartog, R., Karlsson, A., Kaltenegger, L., et al. Configurations for planet finding. SPIE 5487, 2004. Dubovitsky, S., Lay, O.P. Architecture selection and optimization for planet finding interferometers. SPIE 5491, 2004. Eiroa, C., Fridlund, M., Kaltenegger, L. The DARWIN target list: observational properties of the G-type stars. ESA SP 539, 403–407, 2003. Fischer, D., Valenti, J.A. Metallicities of stars with extrasolar planets. ASP Conference Series 294, 117–128, 2003. Gondoin, P., Absil, O., den Hartog, R., et al. Darwin-GENIE: a nulling instrument at the VLTI. SPIE 5491, 2004. Hinz, P.M., Angel, J., Roger, P., Hoffmann, W.F., et al. First results of nulling interferometry with the Multiple-Mirror Telescope. SPIE 3350, 439–447, 1998. Karlsson, A.L., Wallner, O., Perdigues Armengol, J.M. Three telescope nuller based on multibeam injection into single-mode waveguide. in: New Frontiers in Stellar InterferometryProceedings of SPIE, vol. 5491. The International Society for Optical Engineering, Bellingham, WA, p. 831, 2004. Kaltenegger, L. Search for extraterrestrial planets: the DARWIN mission. Target Stars and Array Architectures, PhD Thesis, Karl Franzens University Graz, 2004. Kaltenegger, L., Karlsson, A. Requirements of the stellar rejection for the Darwin mission. SPIE 5491, 2004. Kaltenegger, L., Fridlund, M. Multiple planets, 2004. Available from:. Karlsson, A., Kaltenegger, L. The technology of Darwin, in: Proc Towards Other Earths: DARWIN/TPF and the Search for Extrasolar Terrestrial Planets, ESA SP, vol. 539, 2003. Kasting, J.F., Whitmire, D.P., Reynolds, R.T. Habitable zones around main sequence stars. Icarus 101, 108, 1993. Leger, A., Mariotti, J.M., Mennesson, B., Ollivier, et al. Could we search for primitive life on extrasolar planets in the near future? The DARWIN project. Icarus 123, 249–255, 1996.
1122
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
Lunine, J.I. Special feature: the occurrence of Jovian planets and the habitability of planetary system. PNAS 98, 809, 2001. Marcy, G.W., Butler, R.P. A planet orbiting 47 Ursae Majoris. ApJ 464, L147, 1996. Marcy, G.W., Butler, R.P., Fischer, D., et al. Properties of extrasolar planets. ASP Conference Series 294, 1–16, 2003. Mazeh, T., Naef, D., Torres, G., et al. The spectroscopic orbit of the planetary companion transiting HD 209458. ApJ 532, L55, 2000. Mayor, M., Queloz, D. A Jupiter–Mass companion to a solar-type star. Nature 378, 355, 1995. Mennesson, B., Mariotti, J.M. Array configuration for a space infrared nulling interferometer dedicated to the search for Earth-like planets. Icarus 128, 202–212, 1997.
Ollivier, M., Selsis, F., Leger, A. Finding and characterizing extrasolar planets at (thermal) infrared wavelengths. SP 539, 241, 2003. Owen, T. The search for early forms of life in other planetary systems: future possibilities afforded by spectroscopic techniques, strategies for the search of life in the universe. Strategies for the Search for Life in the Universe, 177–185, 1980. Perryman, M.A.C. Extra-solar Planets. Rep. Prog. Phys., 69, 2000. Selsis. The Atmosphere of Terrestrial Exoplanets: Detection and Characterization, Extrasolar Planets: Today and Tomorrow. ASP Conference Proceedings 321, 2004. Serabyn, E. Nulling interferometry: symmetry requirements and experimental results. SPIE 4006, 2000. Velusamy, T., Angel, R.P., Eatchel, A., et al. Single and double Bracewell nulling interferometer in space. ESA SP 539, 2003.
The Darwin mission: Search for extra-solar planets L. Kaltenegger a
a,*
, M. Fridlund
b
Science Payload and Advanced Concepts Office, European Space Agency ESTEC, P.O. Box 299, NL-2200AG Noordwijk, The Netherlands Research and Scientific Support Department, European Space Agency ESTEC, P.O. Box 299, NL-2200AG Noordwijk, The Netherlands
b
Received 14 January 2005; received in revised form 19 May 2005; accepted 19 May 2005
Abstract The direct detection of an Earth-like planet close to its parent star is challenging because the signal detected from the parent star is between 109 and 106 times brighter than the signal of a planet in the visual and IR respectively. Future space based missions like Darwin and TPF-I concentrate on the mid-IR region between 6 and 20 lm, a region that contains the CO2, H2O, O3 spectral features of biomarkers in EarthÕs atmosphere. The InfraRed Space Interferometer Darwin is an integral part of ESAs Cosmic Vision 2020 plan, intended for a launch towards the middle of the next decade. It has been the subject of a feasibility study and is now undergoing technological development. It is focused on the search for, and characterization of Earth-like planets orbiting other stars. A secondary objective is to carry out imaging of astrophysical objects with unprecedented spatial resolution. The implementation is based on the new technique of Ônulling interferometryÕ. New designs have been developed that will be implemented on four spacecrafts and search for planets around a minimum of 165 stars within the mission lifetime. 2005 COSPAR. Published by Elsevier Ltd. All rights reserved. Keywords: Darwin; Extrasolar planet search; Biomarkers; Nulling interferometry; Darwin targets stars; TPF-I
1. Introduction The InfraRed Space Interferometer Darwin is a major element in the Cosmic Vision 2020 program of the European Space Agency. Darwin has the explicit purpose of detecting other Earth-like worlds, analyze their characteristics, determine the composition of their atmospheres and investigate their capability to sustain life as we know it. The closing years of the 20th century have allowed us, for the first time, to seriously discuss interferometric instruments deployed in space achieving unprecedented spatial resolution. These missions will lead to new astrophysics. Especially – and this is the greatest challenge – * Corresponding author. Present address: CfA, Astronomy Department, 60 Garden Street, Cambridge, MA 02138, United States. Tel.: +1 617 495 7158. E-mail address: [email protected] (L. Kaltenegger).
we expect to be able to carry out the first detailed study of terrestrial exoplanets (defined as planets similar to our own Earth as what concerns size and mass, and orbiting other stars than our Sun) as well as comparative planetology based on the diversity of planets detected. The detection and study of the latter promises to open a new era in science and will affect a broad spectrum of science and technology. We can now confidently expect the first results from space based interferometers within 10 years. Sophisticated instruments will follow in short order. Analysis of the planetary light requires that the stellar light is suppressed to a high degree. This is done by a technique called nulling interferometry, in essence this means that achromatic phase shifts are applied to the beams collected by individual telescopes before recombination such that the on-axis light, i.e., stellar light, is cancelled by destructive interference, while the much weaker planetary light emitted at a certain off axis angle interferes constructively.
0273-1177/$30 2005 COSPAR. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.asr.2005.05.061
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
The Darwin mission is implemented on four free flying spacecraft including one beam-combining spacecraft. The beam combiner and the telescope spacecraft fly in one plane with each telescope spacecraft at the same distance from the beam combiner. The resolution of the interferometer is adjusted by changing the distance between the telescope spacecrafts. A similar activity has been taking place in the United States within the context of NASAs Origins program. A science collaboration has already been established.
2. Extrasolar planets The last years have seen the detection of planets beyond our solar system finally becoming a fait accompli. The techniques utilized so far have been indirect. Radial velocity relies on measuring the reflex motion of the parent star, with respect to the common center of mass of the star–planet system. The transit method measures the reduction in flux when a planet crosses the line of sight between earth and its host star. The first detection of a planet outside the Solar System using radial velocity method was reported by Mayor and Queloz (1995) for the Solar type star 51 Peg. This was quickly followed by the Lick group who reported planets around 70 Vir and 47 Uma (Marcy and Butler, 1996; Butler and Marcy, 1996). It appears unlikely, however, that this method could be used to infer the presence of Earth-type planets. The current precision in radial velocity is 3 m s 1, while the Earth at 1 AU from the Sun would require 0.1 m s 1. Acoustic pressure-mode oscillations in solar-type stars have amplitudes of 0.5–1 m s 1, and will thus make it essentially impossible to detect a deflection of an Earth simile. The planets so far found with the radial velocity method are objects more akin to the planet Jupiter than something like our own Earth. Radial velocity searches for planets are strongly biased toward planets with large mass and short orbital period because of their easy detectability. As time passes longer and longer periods are being picked up, and there are now a number of confirmed planets in orbits with periods of several years. A few of the objects are likely Brown Dwarfs. The radial velocity method provides us with a lower limit to the mass and the orbital radius. If the inclination of the planetary orbit is not known, there is no way to determine the planets absolute mass, actual size or composition – without other data. One trend that has been seen in radial velocity star surveys is that a plot of the number of detected objects vs. M sin(i) increases strongly towards M sin(i) < 1 Mjup. An unknown process would have to bias our detection towards only picking up systems seen face on (Fischer and Valenti, 2003; Marcy et al., 2003). In the case of the first observed occultation, the planet orbits the star HD 209458 every 3.52 days at a distance of
1115
about 0.05 AU. The occultation lasts about 2.5 h and from this observation, the inclination is found to be 87.1. Charbonneau et al. (2000) derive a planetary mass of 0.63 MJup and a planetary radius of 1.27 RJup. This can be done since the orbital radius is well known from the radial velocity measurements, and the stellar radius is known to good accuracy from stellar evolution theory. The actual shape of the light curve during the occultation (Mazeh et al., 2000), constrains the planetary radius and orbital inclination to a very high degree. The average density of the planet turns out to be only half that of the major gaseous giant planets in our own Solar System immediately ruling out that the planet is a rocky, terrestrial body (super-Earth) that could have formed close to its current location since such a planet would be significantly smaller than 1.27 RJup. The planet is thus a gas giant. Being physically larger than Jupiter but with a lower mass is caused by its proximity to its primary which heats it to a surface temperature of 1200 K. Such temperatures would, however, only affect the outer 1% of the planet, and the large diameter immediately says something about its evolution. As pointed out by Lunine (2001), what is happening is that the flux from the star retards the cooling of the planetary interior. A giant planet formed in isolation would cool in a brief time (106 years), and thus it also shrinks rapidly from its original distended state. For a planet in very close proximity to a star such as is the case for HD 209458b, the atmospheric temperature profile is flattened and the rate by which heat can be transported outwards from the interior is reduced and the contraction will be retarded. Detailed models (Burrows et al., 2000) show excellent agreement with the planetary radius at its current age of 4–7 billion years (the age being determined from stellar evolution theory). It can also be shown in these models, that the planet must have arrived at an orbital radius of 0.05 AU within at most a few tens of millions of years after formation. Otherwise it would take longer than the present age of the Universe for the external heat to diffuse inwards far enough to expand the radius to the observed value. The observation of a single occultation thus shows that the so called Ôhot JupitersÕ either form in place or migrate inwards within at most a few · 107 years (Lunine, 2001; Burrows et al., 2000). Another indirect method is to obtain astrometric data and thus track a starÕs path across the sky, measuring the wobble introduced by the rotation around the common center of mass of the star–planet system. The European Space AgencyÕs GAIA mission promises large statistical surveys of massive planets (Perryman, 2000). 2.1. Direct detection of terrestrial planets All of the above mentioned methods will continue to refine our knowledge about planetary systems. Unfortunately they are restricted towards the indirect detection
1116
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
of relatively massive planets. Space missions (e.g., COROT, Eddington, Kepler) designed to study occultation will eventually pick up lower (Earth-size) mass bodies, but relatively little information will be gathered in this fashion. The most important datum will be the frequency of Earth-like planets. The main problem in the direct detection of an exoplanet of a size comparable to our own Earth – and located at a similar distance from its own star – is one involving contrast and dynamical range. A central star like our Sun (G2V) outshines an Earth-like planet in the visual wavelength range by a factor of at least 109. Going to the mid-infrared alleviates this problem, because the planets thermal emission peaks (a terrestrial planet is in this case defined as one of roughly the same size as the Earth and of the same surface temperature – ca. 270 K – thus having its peak emission at 10 lm). Even at these wavelengths, the contrast is more than a factor of 106. The star and planet will be very near each other on the sky, and we need to devise a way of extinguishing the light from the star. Different coronographic methods (in space) have been evaluated, and albeit having the capability of achieving the scientific objective of detecting the exoplanet, these methods do not lend themselves to a large enough search space (unless the telescope is extremely large). Lately it has been suggested to fly coronographic systems operating in the visual wavelength range in space. A 5–10 m diameter monolithic telescope could then suffice but would investigate a smaller search space than the interferometer proposed.
3. Nulling interferometry Infrared nulling interferometry seems to be the preferred instrumental concept for the implementation of a mission observing a large sample of at least 165 stars, for which technological readiness can be achieved in a time frame compatible with a launch during the second half of the next decade. A nulling interferometer extinguishes light from an on-axis source. The basic concept is to sample the incoming wavefront from the star and its planet with several telescopes that individually do not resolve the system. Nulling interferometry differs from ÔnormalÕ imaging interferometry (Michelson interferometry) in that one attempts to obtain a dark fringe (or fringe pattern) in the center of the field, by introducing phase shifts in the light paths of one or several of the interferometer arms. By keeping a star in the center of the image plane, coronography is realized without the presence of a physical mask. In its simplest form, a two-telescope nulling interferometer introduces a 180 phase-shift between the two apertures, resulting in destructive interference along the line of sight. At the same time, light at
small off-axis angles from the line of sight will experience constructive interference, thus allowing a faint object close to a bright star to be discernible. The observable area around the central null depends on the separation of the telescopes – the baseline. The star can be centered on the deep central null. Specific off-axis locations like the position of an earth-like planet orbiting the host star can be constructively interfered by adjusting the distances between telescopes. The output of the system can be described by an angular transmission map (TM) featuring interference fringes, with a sharp null (destructive interfered area) in the center of the map. The stellar signal is nulled out only on the optical axis. A leakage of photons out of the central null exists because the star has a finite photospheric disk. That leakage is a very important noise source. Information about the distribution of planets in the target system can be recovered by modulation of that signal. No spatial information is extracted in a single exposure. Rotation modulates the interferometer output intensity as a planet passes in and out of the dark fringes. From the intensity and actual pattern of this modulation one can derive the planetÕs parameters see e.g., Ollivier et al. (2003) and Kaltenegger and Fridlund (Kaltenegger and Fridlund, 2004). The use of more telescopes achieves a symmetric pattern around the star, with a deep central null placed on the starÕs position. The actual shape and transmission properties of the pattern are a function of the number of telescopes, configuration, and the distance between the telescopes (Mennesson and Mariotti, 1997; Absil, 2001; Kaltenegger, 2004). In practice, the central null of the transmission map will be degraded by amplitude and optical path differences between the interferometer arms what will lead to a level of noise that sets a limit for the achievable null depth. The properties of the Bracewell array in the presence of phase and amplitude perturbations have been discussed by Serabyn (2000). The use of a single Bracewell as nulling interferometer for planet detection is limited by the impossibility to rapidly modulate the planet signal against disturbances such as stellar leakage and the background. For a system at 10 pc observed at 12 lm the local zodiacal cloud signal is the biggest noise source, followed by the stellar leakage and the signal of the exo-zodiacal dust disk. Here, we assume an exozodiacal dust disk similar to our own system. In order to separate out the signal of the planet from this background one needs to modulate the planetÕs signal and that from the zodiacal dust at different frequencies. The planetary signal is separated out from that of any extrasolar zodiacal dust because of its temporal and spectroscopic behavior (Bracewell and McPhie, 1979; Leger et al., 1996; Mennesson and Mariotti, 1997; Absil, 2001) due to modulation. The modulation technique, envisioned is internal modulation. It is based on
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
combining the outputs of various sub-arrays under a variable phase shift (see e.g., Absil et al., 2003). Alternative modulation techniques are under investigation (dÕArcio et al., 2004) especially in the context of GENIE, a proposed ground-based pre-cursor for DARWIN at the VLTI. The influence of exo-zodiacal clouds is an important constraint for planet detection with DARWIN. This issue will be addressed by Ground-based European Nulling Interferometry Experiment (GENIE) (Gondoin et al., 2004). The prime objective of GENIE is to gain experience with the design, construction and operation of a nulling interferometer, as preparation for the DARWIN mission. GENIE will be particularly sensitive to warm circumstellar dust as it operates at mid-infrared wavelengths and thus provides the opportunity to investigate the properties of the DARWIN target stars (Absil et al., 2003; Kaltenegger and Karlsson, 2004). Fig. 1 shows that the detected signal is governed by the contribution of background noise to the interferometer output signal (where star leakage and EZ referee to the residual signal from the star and the exo-zodiacal dust disk respectively, and LZ and thermal referee to the flux from our local zodiacal cloud and the telescopes at 30, respectively). Fig. 1 shows the results for the Three Telescope Nuller (TTN) architecture (Karlsson et al., 2004) (see explanation in Section 4). The output beam comprises all information from a specific set of sources on the sky (consisting of star, planet(s) and zodiacal dust surrounding the target star), as well as the background. The local zodiacal cloud (LZ) provides the foreground through which DARWIN will observe. Because zodiacal background is diffuse, it cannot be cancelled by nulling interferometry. The stellar leakage is dominating for short wavelengths, while the thermal background level due to the emission of the optics, can approach that of the zodiacal dust for temperatures above 40 K at long wavelengths. However, the LZ dust disk remains the biggest noise factor, for most of the wavelength band if we assume that the exo-zodiacal
Fig. 1. Background and earth-sized planet for a G star system at 15 pc as a function of wavelength (Kaltenegger, 2004). The graphs show the modulated signal as obtained for a triangular TTN configuration (Karlsson et al., 2004), including detector characteristics and fiber coupling effects. The planet is shown on a transmission maximum.
1117
cloud (EZ) is similar to our own LZ cloud. Note that large coherent structures such as wakes and clumps behind planets can masquerade as planets. Those structures could also serve as markers for the presence of planets, if their location in respect to a planet were well understood. As exo-zodical clouds are not uniform, a planet must be detected against a non-flat field of corrugations. Structures in our own cloud have roughly <0.1% of the amplitude of the total cloud brightness (Backman et al., 1998) and are thus no source of confusion. At levels less than 1% of the total brightness in a simulated DARWIN observation scenario, the structured emission is not a significant contributor to the noise budget. In an ideal scenario with only a single planet around its host star and no other disturbing sources, such as extrasolar zodiacal dust in the target system, the detection of a positive flux would imply that a planet is present, if the star is well and truly Ônulled outÕ. In real observations, several factors affect the signal to noise in a detrimental way. This has been analyzed by different groups e.g., Velusamy et al. (2003) and Kaltenegger and Karlsson (2004).
4. Architecture design The original DARWIN mission concept was optimized for stellar rejection to focus on observing the closest stars for planetary companions. This led to a baseline concept of a free flying configuration with six collector telescopes of 1.5 m diameter and a central beam combiner. A number of alternative mission architectures have been evaluated on the basis of interferometer response as a function of wavelength, achievable modulation efficiency, number of telescopes and starlight rejection capabilities (Kaltenegger, 2004; den Hartog et al., 2004; Dubovitsky and Lay, 2004). These analyses show that the starlight rejection criteria can be relaxed while still maintaining a target sample of 165 stars. The integration time for close by stars increases, but the overall mission performance does not degrade significantly. The leakage term and the resulting null degradation depend strongly on the baseline used. Optimizing the baseline of the interferometer for each stellar target system will minimize the leakage, while a fixed baseline as would be used for an interferometer implemented on a structure would lead to a high level of leakage for most of the target systems (distances vary between 3 and 25 pc) and have a variation of optimized baselines. Accordingly, candidate Darwin and TPF configurations use three (Karlsson et al., 2004) or four telescopes of 3.5 m diameter, reducing complexity and cost of the mission. Three telescopes is the minimum number of telescopes needed for an interferometer mission that
1118
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
uses rapid signal modulation to detect a planet in the high background noise. The baseline mission duration is 5 years, extendable to 10 years in an L2 orbit. Operating in the infrared band requires that all optical components are cooled to roughly 40 K, this is achieved by passive cooling. Only the detector requires active cooling. The first practical demonstration of nulling was undertaken in February 1998 (Hinz et al., 1998). Using the Multiple Mirror telescope on Mount Hopkins, Arizona, Angel and his team were able to cancel out the image of a star: a-Orionis. The nulled image had a peak in intensity of 4.0% and the total integrated flux of 6.0% of the constructive image.
5. Biomarkers The spectral range between 6 and 18 lm and preferably 5–20 lm has been selected for DarwinÕs extra-terrestrial planetary research since the ratio between planetary and stellar light is improved by roughly three orders of magnitude compared to the visible spectrum and since absorption bands of important tracers of life (H2O, O3 and CO2) are present in this range (Ollivier these proceedings). A major goal of the mission is not only to detect terrestrial exoplanets, but also to investigate if the conditions on the planet in question would allow life as we know it to exist and indeed if it already exists. In order to do so we need to define what life is, and how life as we understand it interacts with its environment in an observable way. In the mid-IR the classical signatures of biological activity on Earth are the 9.6 lm O3 band, the 15 lm CO2 band, and either the 6.3 lm H2O band or its rotational band that extends from 12 lm out into the microwave region or both. In the same spectral region, the 7.7 lm band of CH4 is a potential biomarker for early-Earth type planets. The 9.6 lm O3 band is highly saturated and is thus a poor quantitative indicator, but an excellent qualitative indicator for the existence of even traces of O2. Observations from 8 to 12 lm of the H2O continuum allow estimates of the surface temperature of Earth-like planets. However, one should note that the atmosphere of planets that are warmer than about 310 K will be opaque in this region because of continuum absorption by water vapor. While the existence of H2O in liquid state on the surface of a planet is considered essential for the development of life, it is not by itself a bio-indicator. The presence of CO2 in the atmosphere of all terrestrial planets in our Solar System is of course inconclusive as what concerns signs of biological activity but provides information about the physical state of the atmosphere. Taken together with molecular oxygen, abundant CH4 can indicate biological processes, although depending
on the degree of oxidation of a planetÕs crust and upper mantel also non-biological origins can under certain circumstances produce large amounts of CH4. With a spectral resolution, Dk = 20–40, we can detect all of the features mentioned above. We have to remember, that this refers to our Earth as it is today. The oxygen level in our atmosphere has been relatively high only for the last 20–30% of the EarthÕs history. Although, it is now more or less generally accepted that life arose on the Earth immediately after the era of bombardment, i.e., 3.8 · 109 years ago, it remained in the sea, and at a relatively simple level until just about 600 million years ago. Life before the rise of oxygen in the atmosphere around 2.5 Gyr ago was dominated by Methane producing species and if we want to define our Ôremote sensingÕ criterion such that life is indicated by a disturbance of the equilibrium of a terrestrial planets atmosphere, we need to also take this evolutionary aspect into account. In order to interpret spectroscopic data in the context of possible biological activity we also need to know the planetÕs size and temperature. In the thermal part of the spectrum, the shape gives a measure of the temperature of the object examined. The mid-IR spectra can give the planetÕs albedo, the temperature of the observable emitting regions and thus the planetÕs size. If all life on the Earth were removed suddenly – bacteria, green plants, etc., all of the free oxygen in the EarthÕs atmosphere would disappear in the geologically short time of 4 million years. The atmosphere of the Earth is out of equilibrium as is evidenced by a comparison with models or with the situation in the other terrestrial planets in the Solar system – Mars and Venus. This disequilibrium is caused by the living things on our planet. Previous to life being dominated by Oxygen generating species, the atmosphere of the early Earth was out of equilibrium by Methane, CH4.
6. Life and habitability The search for life requires a general definition of life. Life contains information; life is self-replicating; life evolves; and life influences its environment. We are attempting to detect these attributes of life, through remote sensing, at interstellar distances. However, no clearcut observables can be indicated – particularly such that can be detected over interstellar distances. One should, however, remember, that based on the two first criteria, we would also determine that a computer virus is alive. We have to expand our criteria. Faced with this problem, space agencies currently base their development of missions on what life on our own planet looks like. To define search criteria, we presume that life has a carbon-based structure not too dissimilar from our own. Owen (1980) has argued that carbon chemistry with water as a solvent is the most
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
likely base of life. Carbon is a highly abundant element and exhibits a remarkable ability to form a host of highly complex molecules. Arguments in favor of water as a solvent are that it is highly abundant, an excellent solvent and remains liquid over a broad temperature range. The temperature range is high enough for chemical reactions to occur rapidly but also allows large molecules to form. Specific conformation of macromolecules can form when in solution in water due to the solventÕs attraction of hydrophilic groups and repulsion of their hydrophobic ones. These conformations allow very specific chemical reactions to build reproducible complex structures. Water dissociates to form oxygen and ozone which can protect liquid water from further dissociation, provided there is a cold trap in the atmosphere on the hypothetic planet to confine water to lower altitudes. Basing the search for life on the carbon chemistry assumption it is then possible to establish criteria for habitable planets in terms of their size and distance from their stars. Gases in their atmosphere such as oxygen, methane and water vapor would indicate in combination the presence of life (see Fig. 2). Based on the history of our own planet, atmospheric features can provide clues of possible life forms for at least 2 billion years. This is more than 107 times longer than radio signals reveal the presence of an advanced civilization on our planet. The criterion on temperature defines a habitable zone HZ (Kasting et al., 1993) around each individual star. The HZ around different stellar types will vary as it only depends on the spectral energy distribution of the star. The HZ around F stars is larger and occurs farther out, the HZ for K and M stars is smaller and situated farther in than around our Sun. High ultraviolet fluxes could be a problem for life around F stars. For stars later than about K5 spectral class the HZ falls within the 4.6 Gyr tidal locking radius of the star. The planetary companions may rotate synchronously with their orbit, resulting in a freezing out of the atmosphere on the dark side of the planet. The temperature difference between the bright and the dark side of the planet could generate winds that efficiently reduce the temperature gradient on the planet if the atmosphere were thick enough. A thick CO2 atmosphere would very likely cover bio-signatures like O3, what makes M stars secondary target because biomarkers that indicate the existence of life as we know it would not be detected. Strictly speaking, the surface temperature of a planet will depend not only on the energy input from the primary, but also on the atmospheric pressure and composition. Our own Earth, for instance, would be significantly colder without its green house effect, caused by CO2 and CH4. Since we a priori have no idea about the presence and/or composition of any eventual atmospheres, we will have to use this criterion with some care in an individual case (Selsis, 2004).
1119
Fig. 2. Thermal infrared spectra of Venus, Earth and Mars. The 15 lm CO2 band is seen in all three planets. Earth also shows evidence of O3 and H2O (ESA, http://sci.esa.int/science-e-media/img/c8/planetbig.jpg, 01.10.2004).
A planet with O3 and H2O and CO2 absorption bands in its spectrum that lies within the HZ is the main target of a mission that searches for habitable planets. We point out the possibility that life exists on planets that do not show H2O and O3, if the production of O2 by photosynthesis is not able to overcome the oxygen sinks, e.g., Mars-like planets have small O2-sinks and Venuslike planets large abiotic O2 sources (Kasting et al., 1993).
1120
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
7. Target star catalogue An early target list of stars, mainly from the Gliese and Jahreiss catalogue of nearby stars (1991) was compiled by A. Leger and M. Ollivier for the Darwin feasibility study for F, G and K stars. In preparation of the Darwin mission we (Kaltenegger, 2004) established an updated catalogue including new information and an extended star sample of stellar systems of nearby FGKM stars that are potentially habitable to complex life. The list was created from the Hipparcos catalogue by examining the information on distance, stellar variability, multiplicity, location and spectral classification as well as the interferometerÕs limitation. Four selection criteria were adopted: (i) FGKM stars within a distance of 25 pc; (ii) main sequence stars nat-
ure based on the B-V index; (iii) stars within a cone of aperture ±45 of the ecliptic plane; (iv) Stars with magnitudes smaller than 12 in V. Limitations imposed by the instrument are e.g., the 45 observation cone in anti-sun direction, a minimum distance to a close binary companion that cannot be simultaneously nulled while searching for a planet and a minimum observation time of the star in the field of view of the instrument. The 45 cone in the selection criteria is given by constraints of the sunshields on each free flyer. The shields will be designed to permit a ±45 cone of observation in the anti-sun direction. That also influences the maximum time during which a target can be observed. The selection criteria for the updated DARWIN stars from the HIPPARCOS catalogue lead to a total of 807 stars. The current analysis results in a prime target list of
Fig. 3. Spatial distribution of all prime DARWIN target stars (FGKM) (Kaltenegger, 2004).
Fig. 4. Spatial distribution of FGK prime DARWIN target stars (Kaltenegger, 2004).
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
1121
Darwin has the explicit purpose of detecting other Earth-like worlds, analyze their characteristics, determine the composition of their atmospheres and – investigate their capability to sustain life as we know it, an exciting challenge that we are prepared to answer during the second half of the next decade.
References
Fig. 5. Distribution of DARWIN prime targets over distance (Kaltenegger, 2004).
285 target stars of stellar type F, G, K and selected M stars excluding multiple systems and active stars, consistent with the selection criteria within 25 pc of our sun (see Figs. 3–5). For the G type stars the properties of the stellar targets were revised (Eiroa et al., 2003) through consultation of existing data archives. Note that the HIPPARCOS catalogue is not complete to the 12th magnitude. As Darwin concentrates on brighter/closer star the completeness of the HIPPARCOS sample is an issue for densely populated areas. Thus, the number of stars quoted is a first analysis. Further data on nearby stars will be gathered and the target list will be updated once new data is available. The free flyer array architecture can be optimized to detect planets in the Habitable zone around the selected target stars. This is a big advantage of a free flying array over a structurally connected interferometer.
8. Conclusion The Darwin model mission represents today a design that could answer one of the longest lasting scientific questions – if we are alone in the Universe. The feasibility study has demonstrated the maturity of the existing technology (Karlsson and Kaltenegger, 2003). The design assessment leads to a design consisting of four free flying spacecrafts that can observe a minimum of 165 stars from the DARWIN target star list (Kaltenegger, 2004). Infrared nulling interferometry emerges as the preferred instrumental concept for the implementation of a mission observing a large sample of at least 165 stars, for which technological readiness can be achieved in a time frame compatible with a launch during the second half of the next decade.
Absil, O. Nulling Interferometry with IRSI Darwin: Detection and Characterization of Earth-like Exoplanets, Master Thesis, Universite´ de Lie`ge, 2001. Absil, O., Karlsson, A., Kaltenegger, L. Inherent modulation: a fast chopping method for nulling interferometry. Proc. SPIE 4852, 431– 442, 2003. Backman, D.E., Caroff, L.J., Sanford, S.A. et al. (Eds.), Proceedings of the Exozodiacal Dust Workshop (NASA/CP-1998-10155), 1998. Bracewell, R.N., McPhie, R.H. Searching for nonsolar planets. Icarus 38, 136–147, 1979. Butler, R.P., Marcy, G.W. A planet orbiting 47 Ursae Majoris. ApJ 464, L153, 1996. Burrows, A., Guillot, T., Hubbard, W.B., Marley, M., et al. On the radii of close-in giant planets. ApJ 534, L97, 2000. Charbonneau, D., Brown, T.M., Latham, D.W., Mayor, M. Detection of planetary transits across a Sun-like star. ApJ 529, L45, 2000. dÕArcio, L., Karlsson, A., Godoin, P. Use of OPD modulation techniques in nulling interferometry. SPIE 5491, 2004. den Hartog, R., Karlsson, A., Kaltenegger, L., et al. Configurations for planet finding. SPIE 5487, 2004. Dubovitsky, S., Lay, O.P. Architecture selection and optimization for planet finding interferometers. SPIE 5491, 2004. Eiroa, C., Fridlund, M., Kaltenegger, L. The DARWIN target list: observational properties of the G-type stars. ESA SP 539, 403–407, 2003. Fischer, D., Valenti, J.A. Metallicities of stars with extrasolar planets. ASP Conference Series 294, 117–128, 2003. Gondoin, P., Absil, O., den Hartog, R., et al. Darwin-GENIE: a nulling instrument at the VLTI. SPIE 5491, 2004. Hinz, P.M., Angel, J., Roger, P., Hoffmann, W.F., et al. First results of nulling interferometry with the Multiple-Mirror Telescope. SPIE 3350, 439–447, 1998. Karlsson, A.L., Wallner, O., Perdigues Armengol, J.M. Three telescope nuller based on multibeam injection into single-mode waveguide. in: New Frontiers in Stellar InterferometryProceedings of SPIE, vol. 5491. The International Society for Optical Engineering, Bellingham, WA, p. 831, 2004. Kaltenegger, L. Search for extraterrestrial planets: the DARWIN mission. Target Stars and Array Architectures, PhD Thesis, Karl Franzens University Graz, 2004. Kaltenegger, L., Karlsson, A. Requirements of the stellar rejection for the Darwin mission. SPIE 5491, 2004. Kaltenegger, L., Fridlund, M. Multiple planets, 2004. Available from:
1122
L. Kaltenegger, M. Fridlund / Advances in Space Research 36 (2005) 1114–1122
Lunine, J.I. Special feature: the occurrence of Jovian planets and the habitability of planetary system. PNAS 98, 809, 2001. Marcy, G.W., Butler, R.P. A planet orbiting 47 Ursae Majoris. ApJ 464, L147, 1996. Marcy, G.W., Butler, R.P., Fischer, D., et al. Properties of extrasolar planets. ASP Conference Series 294, 1–16, 2003. Mazeh, T., Naef, D., Torres, G., et al. The spectroscopic orbit of the planetary companion transiting HD 209458. ApJ 532, L55, 2000. Mayor, M., Queloz, D. A Jupiter–Mass companion to a solar-type star. Nature 378, 355, 1995. Mennesson, B., Mariotti, J.M. Array configuration for a space infrared nulling interferometer dedicated to the search for Earth-like planets. Icarus 128, 202–212, 1997.
Ollivier, M., Selsis, F., Leger, A. Finding and characterizing extrasolar planets at (thermal) infrared wavelengths. SP 539, 241, 2003. Owen, T. The search for early forms of life in other planetary systems: future possibilities afforded by spectroscopic techniques, strategies for the search of life in the universe. Strategies for the Search for Life in the Universe, 177–185, 1980. Perryman, M.A.C. Extra-solar Planets. Rep. Prog. Phys., 69, 2000. Selsis. The Atmosphere of Terrestrial Exoplanets: Detection and Characterization, Extrasolar Planets: Today and Tomorrow. ASP Conference Proceedings 321, 2004. Serabyn, E. Nulling interferometry: symmetry requirements and experimental results. SPIE 4006, 2000. Velusamy, T., Angel, R.P., Eatchel, A., et al. Single and double Bracewell nulling interferometer in space. ESA SP 539, 2003.