La Palma International Time Proposal:
Planetary Systems: their formation and properties
(Shortened version)
EXPORT: EXo-Planetary Observational Research Team
MEMBERS:
Carlos Eiroa (1), Alan J. Penny (2), Antxon Alberdi (3), Andrew Cameron (4), John K. Davies (5), Roger Ferlet (6), Francisco Garzon (7), Carol Grady (8), Allan W. Harris (9), Keith Horne (4), Luis F. Miranda (3), Benjamin Montesinos (3,10), Rene D. Oudmaijer (13), Javier Palacios (1), Andreas Quirrenbach (11), Heike Rauer (9), Jean Schneider (14), Paul Wesselius (12), Dolf de Winter(7)
INSTITUTIONS:
(1) Dpto. Fisica Teorica. Universidad Autonoma de Madrid. Spain (2) Rutherford Appleton Laboratory, Oxon, United Kingdom (3) Instituto de Astrofisica de Andalucia, CSIC, Granada, Spain (4) University of St Andrews, United Kingdom (5) Joint Astronomy Center, Hilo, Hawaii, USA (6) Institute d'Astrophysique, CNRS, Paris, France (7) Instituto de Astrofisica de Canarias, Canary Islands, Spain (8) Eureka Scientific, USA (9) Institute of Planetary Exploration, DLR, Berlin, Germany (10) Laboratorio de Astrofisica Espacial y Fisica Fundamental, Madrid, Spain (11) Max-Planck-Institut fuer Extraterrestrische Physik, Garching, Germany (12) Space Research Organization of the Netherlands, Groningen, The Netherlands (13) Imperial College of Science, Blackett Laboratory, London, UK (14) Observatoire de Paris, Meudon, France
Full information of the EXPORT members
Scientific Case
The European Space Agency, ESA, has identified as a cornerstone for its long-term scientific program the launch of an interferometer in Space. One of the options is an infrared interferometer whose main scientific goal is the search for and physical characterization of Earth-like planets.
During the last couple of years an European team -the DARWIN team- composed of scientists of different countries has been working toward the conception and definition of such an infrared interferometer, its technological challenges and its scientific output. As an initiative of the DARWIN team, the first international meeting dedicated to Infrared Interferometry in Space was held last year, March 1996, in Toledo, Spain (Eiroa et al. 1997, proceedings). A main output of discussions within the DARWIN team is the identification of the need to promote precursor studies in both technical and scientific directions, in order to soundly assess the challenges of such an ambitious project. In particular, the team is concious of the scarce knowledge we currently have on planetary systems other than our own.
We would like to remark at this point that ESA began, March 1997, its official studies on an Infrared Space Interferometer establishing an ESA-working group. In addition, the US Space Agency NASA has similar plans as ESA included in their program "ORIGINS", with the definition of the infrared interferometer TPF (Terrestrial-Planet Finder). NASA is also promoting ground-based observational studies on exoplanetary systems by using the NASA Keck observing time and the IRTF.
The ultimate goal of this proposal is to serve as a ground-based scientific precursor of DARWIN. To accomplish this, we are focussing our interest on the study of planetary systems: their formation, evolution, and the search for new planets. These objectives have a scientific merit of their own.
Studies of exoplanetary systems have received considerable impuls since the announcement of a Jupiter-like planet orbiting 51 Peg by Mayor and Queloz (1995). In the interim, further Jupiter-like planets around nearby solar-type stars have been discovered (updated information is available on the web pages created by Schneider and Penny, www.obspm.fr/planets and ast.star.rl.ac.uk/darwin respectively). A main result revealed by all of these discoveries is that our Solar System constitutes just only one particular scenario of planetary systems. In addition, there is evidence that the precursors of planetary systems are widely distributed in spectral type. These discoveries, therefore, strongly support the need of understanding the formation and evolution of planetary systems from a pure observational point of view, and that defines subsequent theoretical modelling, in many different scenarios. In this context, observations of protoplanetary pre-main sequence (PMS) and main-sequence (MS) disks and follow-up studies of stars with exo-planets achieve their sound scientific significance.
In this project, we propose an ambitious observational program addressing the following scientific topics:
i) Formation of planetary systems. This includes a) studies of protoplanetary systems and solid bodies around PMS stars, b) studies of MS disks similar to the beta-Pic system.
ii) Evolved planetary systems. This comprises MS stars in which planets have been discovered. In particular, we will address the recent controversy on the origin of the Doppler shift variations in the 51 Peg spectrum.
iii) Searches for new planetary systems around MS stars, including jovian and Earth-mass planets.
The scientific output will provide a comprehensive observational scenario of planetary systems and relevant constraints for the formation and evolution of planetary system theories.
The scientific team has enough experience with the required observational techniques. Some of its members are in the European DARWIN team and also in the ESA-working group for the infrared interferometer study.
Planetary systems are thought to originate from disk-like structures around young stars. Observations using many different techniques have demonstrated the presence of circumstellar gas and dust disks around solar-type and intermediate mass PMS stars (e.g. Beckwith and Sargent 1996; Mannings and Sargent 1997). Recent radial velocity studies have discovered variations which have been interpreted as the signatures of gas-giant planets orbiting around MS stars (Mayor and Queloz 1995; Marcy and Buttler 1996). Linking the data of PMS disks with those of apparently developed MS planetary systems requires a demonstration that disk formation proceeds via planetesimal formation to the production of planets.
The masses inferred for the PMS disks are on the order of 0.01 to 0.1 solar masses, which is in the range to produce our solar system. Planetary formation starts with the growth of dust particles residing in the disk, where the density is high enough to produce collisions. Calculations show that dust grains will grow by sticky collisions into sizes of the order of 10 km. These planetesimals can begin to interact with each other through gravity and subsequent growth occurs through a runaway accretion phase. If they further fuse, the planetesimals can even grow to sizes of about 2000 km in about 10^5 years, forming planetary embryos (see e.g. Boss 1997). Direct imaging of exo-planetesimals is not feasible with the current and foreseeable technology. Such bodies have substantially less surface area than micron-sized grains and thus do not radiate as efficiently in the infrared. Therefore, we are obliged to go to indirect proofs of their existence.
The beta-Pic system provided the first indication that under favorable conditions the presence of such exo-bodies might be revealed; even more, there are strong indications than the at least 8 Myr-old Beta Pic disk (Crifo et al. ,1997) already harbours one or more giant planets. Beta-Pic presents spectroscopic absorption features due to circumstellar gas at stellar radial velocities together with redshifted components in both the UV and optical spectra (Ferlet et al. 1987, Hobbs 1986). These components have been explained as due to comet-like bodies that evaporate when approaching the star up to 0.5 AU (Beust 1994). The redshifted components are modelled as a superposition of absorption from comae of swarms of star-grazing bodies (Beust et al. 1996), and the multi-year fluctuation of such "infall events" may point to perturbation produced by gas-giant planets. SWS ISO observations show the extraordinary richness of solid-state features in the spectrum of circumstellar disks around young MS stars (Waelkens et al. 1996). The ISO spectra are similar to those found in meteorites and comets and also in beta-Pic (Knacke et al. 1993), which is a further support for the existence of infalling bodies. Therefore, optical observations of the gaseous remmants of the star-grazing bodies combined with ISO results promise to be an extraordinarily powerful tool to probe conditions within a few AU of the star and to study the physical conditions and nature of the dust disks.
Beta-Pic is a MS star. The time scale for dissipation of circumstellar disks and for the formation of large bodies, planets, is shorter than the PMS evolution of such a star. Therefore, we would expect to observe younger stellar systems which already show evidence for planetesimal formation. PMS stars of intermediate mass, the HAeBe stars, are at most a few Myr old. Among them, a group of around 60 objects with spectral type A and large infrared excesses can be considered as beta-Pic progenitors. Within this group, approximately 15 objects exhibit a peculiar photometric behaviour, blueing during deep minima; the prototype of this class is UX Ori (Herbst 1986). Photometric, spectroscopic, and polarimetric data show that the UXORs have a circumstellar disk seen edge-on (Grinin et al. 1994), which seems to contain revolving dust clouds with masses around 10^21 g and radii 8 stellar radius. High and moderate resolution spectroscopy in the optical and UV have demonstrated that these stars have redshifted absorption components, for instance in the NaI D lines, similar to those seen toward beta-Pic, but typically with higher column densities and the infalling material is visible in a wider range of atomic and ionic transitions (e.g. Grinin et al. 1994, Grady et al. 1996a, de Winter 1996). In addition, circumstellar and infall signatures have been found in the line of sight toward a number of near-ZAMS objects and in field A-shell stars (e.g. Grady et al. 1996b, 1997; Dunkin et al. 1997).
Summarizing, there is compelling evidence of the existence of circumstellar disks and solid bodies in the environment of PMS and MS stars, and even perhaps planets, case of beta-Pic. A systematic study of this group of objects, combining optical and ISO data, will reveal important information concerning the properties and evolution of protoplanetary disks and the formation, evolution and properties of those solid bodies, which ultimately most likely lead to the formation of planetary systems.
After many years of trying to find extra-solar planets, the first discovery of a planet orbiting the solar type star 51 Peg was made in 1995 by Mayor and Queloz, as already mentioned; soon thereafter other planets outside our solar system have been found (Butler and Marcy 1996). In the case of 51 Peg, the presence of a planet has recently been questioned by measurements of variations in emission line profiles (Gray 1997) and has been discussed since. It is evident, therefore, that there is an urgent need to find independent, direct arguments to establish the existence of the planet.
In this sense, we will atempt to detect signatures of a possible extended atmosphere of the planet by absorption features in the optical spectrum of the star during transit of its planet. Such a direct measurement of signatures of a planet orbiting 51 Peg could solve the controversy on the origin of the Doppler shift variations in its spectrum. In the following we therefore assume that the Doppler shifts observed are indeed caused by a larger-sized planet and try to find signatures in the star's spectrum to verify this hypothesis. For 51 Peg, a program to search for atmospheric signatures is currently performed in the infrared range with the ISO satellite (Rauer et al. 1996). We propose here to extend this study to optical wavelengths and to other proposed planets which are close companions of their stars.
Although the existence of a large silicate planet cannot be ruled out (Guillot et al. 1996), it is likely that the planets with masses of a few Jupiter masses, M_J, are giant gas planets. First estimates show that for a planet with 0.5 - 2 M_J and a semi-major axis of 0.05 AU (orbital period 4.23 days), as proposed for 51 Peg B, hydrodynamic escape of molecules in its upper atmosphere could be a significant process (Coustenis et al. 1997). The energy input of the solar wind deposited in the exosphere (particle flux of about 10^18 Watts) is capable of tearing away also some of the heavier molecules, such as H2O, CO, CH4, and N2, in addition to H2 and He. High escape velocities allow to cross 3-10 planetary radii, R_J, quickly. Consequently, ions and dissociation products of the escaped molecules should fill the space around the planet, or could even build up an extended tail by interaction with the solar-type wind and radiation field (Schneider et al. 1997).
If 51 Peg B is indeed a large gas giant like Jupiter, it will consist of H, molecular hydrogen, water and carbon (Atreya 1986). In 51 Peg B (T approx. 1300 K) CO is expected to be the dominant carbon-bearing molecule, in contrast to methane in Jupiter (Guillot et al. 1996). Neutral molecules and atoms are expected to be dissociated and/or ionized quickly in the strong solar radiation field. Several vibrational bands of the CO^+ ion, can be observed in the blue optical spectrum in the range from 400 nm to 550 nm. Several additional ionic species have absorption signatures in this range, like CO2^+, N2^+ and CH^+, in addition to a number of neutral dissociation products, like CN, CH, C2, and C3. In the red optical range molecules like H2O^+, O[I] and OH can be observed.
Up to now, 9 stars have been reported with planets around them (see the web pages cited above). In all cases, the implied planet characteristics is that of jovian planets at distances from the parent star ranging from around 0.05 to 2 AU. Although these results are biased because of the detection technique used, it implies that many planetary systems have formed with properties considerably different to those of our own system. In addition to these stars, several other planet candidates have been reported which deserve further confirmation.
There are several observing strategies to search for planets, which are competitive to radial velocity searches. Two of the most promising, which have already even produced some candidates, are microlensing and planet transit searches. Both techniques can be carried out with small 1m class telescopes. The importance of the potential results is that they are sensitive to Earth-like masses and even make statistical studies of the presence of planets.
The microlensing search is based on the alteration produced by a planet in the light curve of a lensing star (see e.g. Ferlet 1997 and references therein). The microlensing events caused by planets have a probability to occur of roughly 10% per event for a star with a Jupiter mass planet, scaling as square root (planet mass). The duration of the events is 1-2 days for a Jupiter, scaling also as square root. This means that a lensing star with an Earth-mass planet produces a lensing peak lasting only hours and with only 1% chance of occurrence per event. The likely outcome of monitoring 30-50 events (about 30-50 events/year are reported by the MACHO project) in an observing season will be the detection of one or several Jupiter-mass planets, and there is a non-negligible chance to detect planets down to Earth masses if they are abundant. We should note that the sensitivity is greatest for planets whose orbit radii match the size of the Einstein ring, about 5 AU for a lensing-star half-way to the galactic center. Sensitivity remains high from roughly 2 to 10 AU, but is low at smaller and larger orbit radii. From the observations it is straightforward to derive the planet/star mass ratio and the projected planet/star separation normalized to the Einstein radius. The technique of planet transits is based on a photometric method, measuring the brightness drop in the light curve of the star produced by the planet. For giant planets in close orbits, like the 51 Peg planet, we have a good chance of assessing the number of such objects and determine their physical size. According to the recent work carried out by Mayor, Marcy and other, the probability of a solar-type star having such a Jupiter-like companion is a few per cent and the probability of having a "correct" inclination of the system is around 0.03 (for an 0.1 AU orbit). The time of transit of the planet in front of the star is one to few hours (considering 3-10 day orbits). The expected brightness drop for a Jupiter-like planet transiting a star is around 1% and the probablity of success in detecting the transit is about 0.1%. These numbers mean that we should look at a few thousand stars, which immediately suggests to observe old open clusters with a photometric precision of around 0.5%.
In addition, the transist method has the unique capability to detect giant (R &> 1 Earth radius) satellites around extrasolar planets (Schneider 1997, submitted). Radial velocity and astrometric searches are definitely insensitive to such objects. The detection by direct imaging of a giant satellite would require at 2 microns a 500 m baseline interferometer. From the ground, it is not possible to detect directly the luminosity drop made by an Earth-sized moon of a giant planet. But its existence would be revealed by the perturbation of the timing of the transits of its parent planet. An Earth-mass moon in an Europa-like orbit makes a shift of approximately 20 minutes between different transists. A further point of interest of transits is that it is the first step toward the investigation of the planet atmosphere by the spectroscopic observation (during transist) of the absorption of the star's light by the planet atmosphere (Schneider, 1994).
Both techniques will potentially provide statistics about the number of stars with planetary systems and also candidates for follow-up studies, as for instance studies of extended planetary atmospheres. These results will be extremely useful for the foreseen space projects and for the study of physical characteristics of very different planetary systems
Proposed Observations
We want to study the transition phase and evolution from protoplanetary disks around PMS stars in which planetesimals have already formed, i.e. UXORs, to disks around MS stars, in which planets might already exist, i.e. beta-Pic analogs. In order to achieve this goal, we intend a systematic study on the basis of a wide sample of stars covering the whole range of evolutionary ages between the PMS and MS stages. Our sample of objects includes UXORs with spectral signatures of infalling bodies, PMS objects with photometric behaviour similar to UXORs, HAe stars close to the ZAMS, A shell stars identified to have accreting gas and beta-Pic/Vega-type stars. Our observations will concretely focus the following topics.
a) Evolution of protoplanetary disks, via near-infrared JHKLM photometry and optical polarimetry combined with ISO and IRAS data.
b) Relevance and properties of solid bodies in the different evolutionary stages. Frequency and duration of infall events. To do that, we need high and intermediate resolution spectroscopy to cover short timescale -days- events, i.e. duration and characteristics of single "comets", as well as to monitor long timescale -months- frequency of infalling bodies. Spectral lines of interest are e.g. CaII H,K, HeI 5876, NaI D, Halpha, etc. High resolution spectroscopy is needed to observe the kinematics and line profiles of the infalling events. Intermediate-resolution spectroscopy also allows us to follow the appearance and disappearance of these events. The simultaneity of both kinds of observation will allow us to estimate abundances and co-variations of different lines/species.
Points a) and b) should be, as far as possible, carried out simultaneously since it is likely that a kind of relationship exists between infalling bodies, their richness and duration, and the polarimetric and photometric behaviour of the stars.
Little is known about extra-solar planets. It is therefore important to cover a wide range of wavelengths to be able to include several possible species and to securely identify absorption features. The observations need to be performed at high resolution to be able to resolve the numerous atomic/molecular absorption features in the stars/planets spectrum and to separate star signatures from atmospheric features. We therefore need the UES spectrograph at the William Herschel
An occultation experiment can be made if the inclination angle, i, of the orbital plane is close to 90 degrees. Measurements show that for 51 Peg B sin i is approximately 1. Thus, an occultation seems likely. From the extent of the exosphere we estimate an occultation to last approximately 3 hAlthough in principle two measurements would be sufficient (one on-transit and one off-transit), following the transit of the planet over 3 h will help in identifying the appearance and disappearance of possible absorption signatures of the planet. Comparison with measurements taken off-transit allows to securely identify absorption features originating from the atmosphere.
In addition to 51 Peg, tau Boo is supposed to harbour a planet with an orbit close to its star (orbital period 3.31 days) that could have an inclination close to 90 degrees. We therefore plan to perform the same observations as for 51 Peg also for tau Boo.
The microlensing search requires R-band exposures taken with the CCD camera of the JKT and we intend to intensively monitoring the light curves of the most promising MACHO alert targets. La Palma has about 4 hours a night during the summer months when the galactic bulge is high enough in the sky. These events are followed by photometry by a number of groups in the Southern Hemisphere, e.g. the PLANET group in Groningen (web page www.astro.rug.nl:80/). Because of La Palma's longitude, these hours fill the gap when the bulge is setting for the South African telescopes and rising for the South American ones. In each measurement, the target will be exposed for 1-2 minutes and that will be repeated 2-3 times. The above sequence will then cycle among an identified set of active MACHO alert targets, with the aim to return to the start of the target list and repeat measurements at least hourly. Ideally, by monitoring 5-10 event lightcurves simultaneously with at least hourly coverage we will have the possibility to detect planets with masses down to that of the Earth. For Jupiter-mass planets, each target must be observed from La Palma at least once or better twice on each available night. In order to increase the time coverage in summer beyond the possibilities of the La Palma international time, we will try to obtain time with the IAC 0.8m telescope on Tenerife.
For the planet transit searches, we plan to observe three well populated open clusters. Given the field of view of the JKT CCD camera we will need some mosaicing to cover the clusters. Since the observations should be carried out during periods of 10-12 days, which is beyond the possibilities of La Palma international time program, we will try to obtain time with the IAC 0.8m telescope on Tenerife. If we assume that 3% of the stars in our clusters have 51_Peg_like planets, we estimate that in a cluster of approx. 2500 stars we will be able to observe 7 51_Peg-like transits. These 7 transits would serve as alerts for further observations.