The First Known Transiting Planet
HD 209458 b was the first planet confirmed using the transit method. The light dimming due to its transit was measured by Charbonneau et al. (2000) and Henry et al. (2000). The fact that it was transiting also allowed other system parameters to be determined: Rp=1.27±0.02 R, i=87.1±0.2 deg. The planet is now very well studied, and these parameters have been refined to Rp=1.359±0.015 R and i=86.71±0.05 deg using a variety of methods, such as longer-duration monitoring (Robichon & Arenou, 2000; Castellano et al., 2000) and spectroscopy observations (Mazeh et al., 2000). Other techniques used for improving parameters have been multi-colour observations (Saurabh et al., 2000, Knutson et al., 2007), simultaneous fitting of transit and radial velocity measurements (Wittenmyer et al., 2005) and improving assumptions of stellar parameters (Torres et al., 2008). Winn et al. (2005) modelled the Rossiter-McLaughlin effect and determined the inclination of the planetary orbit relative to the stellar equator. They found a small but non-zero misalignment of -4.4 ± 1.4 degrees.
Anticipating that such a transiting planet could potentially have a measurable atmosphere, Seager and Sasselov (2000) used models of close-in extrasolar planets to predict absorption signatures that might be seen, such as strong Na I and K I lines in the optical spectrum.
Brown (2001) provided further predictions of the spectra of hot Jupiters, including molecular species, whose features appear in the infrared. They highlighted how important observations are at these wavelengths. Theoretical work was expanded by Hubbard et al. (2001) to consider atmospheric effects such as Rayleigh scattering, cloud scattering and refraction as well as molecular absorption of starlight. They predicted that the planetary radius will be a function of wavelength, with H2O opacity present at infrared wavelengths.
These predictions were soon followed up by the first detection of a planetary atmosphere, using transmission spectroscopy. Using the knowledge of expected spectral features, Charbonneau et al. (2002) used the HST to measure the transit depth around the Na I feature compared to the transit depth in bands outside this feature, to determine whether there was additional atmospheric absorption. They found an excess absorption depth due to the feature of , which confirmed that sodium was present in the planet’s atmosphere. This discovery was very exciting, as it showed that it was possible to obtain information about the atmospheres of transiting planets, not just their existence. The new field of exoplanet atmosphere study began with these observations, and has led to this planet being one of the most studied outside our solar system. Interestingly, the absorption depth was significantly lower than predicted by models, leading to predictions that silicate and iron clouds may reside high in the atmosphere at several mbars (Fortney et al., 2003). Ionisation also occurs around 0.5 mbar, which will weaken the sodium feature.
This observation was then furthered by Sing et al. (2008) who followed this work by using the same technique but varying the band size centred on the sodium line to give an integrated profile of additional absorption depth in the sodium feature vs bandwidth. The line wings were clearly seen, indicating a haze free atmosphere. The new measurements were consistent with the measurements of Charbonneau et al. (2002) in 12, 38 and 100 Angstrom bands around the sodium feature. However, using narrower bands enabled the narrower part of the sodium feature to be measured, and Sing et al. (2008) found that it had an absorption depth of ~0.11%, much higher than the depth found initially in the 12 Angstrom band. This and the observation of broad line wings indicates a comparatively clear atmosphere rather than one with a high-altitude absorber, as in the case of HD 189733 b. This work also analysed the broadband G750L and G430L archive data, and the authors found a rise in absorption depth with decreasing wavelength. Lecavelier des Etangs et al. (2008) explained this blueward rise as Rayleigh scattering due to H molecules. Knowing the abundance of the species responsible for the continuum signature allows its pressure to be determined. Lecavelier des Etangs et al. (2008) found a pressure of ~ 33 mbar and determined a relationship of pressure against altitude.
This does not preclude optical absorbers in other parts of the spectrum, such as TiO and VO. Désert et al. (2008) looked for evidence of these features in the optical HST transmission spectrum. They found excess absorption at 6200-8000 Angstroms but no readily identifiable features, suggesting that such absorbers may still be a possibility.
Once the Na I feature had been detected, the limits of available instruments could be tested even further. Snellen et al. (2008) presented the first ground-based detection of the sodium doublet at high resolution, with the feature partially resolved. The absorption depth measurement was consistent with the observation from Charbonneau et al. (2002) and Sing et al. (2008), and added information about even higher altitudes.
Even more information about a planetary atmosphere can be extracted from transmission spectroscopy, as demonstrated by Sing et al. (2008b). They used altitudes derived from their sodium integrated-averaged absorption profile to convert each region of the profile to a pressure range, and derive characteristic temperatures for each region. This gave the first planetary temperature-pressure profile. Using improved models of the sodium doublet feature, this profile was then refined by Vidal-Madjar et al. (2011, 2011b). These works found an increase in temperature in the upper atmosphere, at very low pressures, indicating a detection of the thermosphere.
The discovery of this first transiting planet also opened up the area of secondary eclipse spectroscopy and photometry, from which huge amounts of physical information can be obtained. These observations are very difficult, due to small signals, and initially unknown timing of the occultation. For this reason, the first few observations were non-detections, only able to place upper limits on emission at 3.6, 2.2 and 2.3 microns (Richardson et al., 2003; 2003b; Brown et al., 2002, Deming et al., 2005). These observations were, however, able to rule out some models that assume total reradiation of stellar heat, suggesting instead that some portion of the heat must be recirculated to the night side.
The first firm detection of an emission feature came from Deming et al. (2005b), who found a brightness temperature of 1130±150 K at 24 microns. Further models were then developed by Fortney et al. (2005), showing that the measurement is consistent with solar metallicity models. However, Burrows et al. (2006) find that even models with 10 times solar metallicity can still be consistent with the data.
Further tentative features have now been identified, including water (Beaulieu et al., 2010, Knutson et al., 2007b, Knutson et al., 2008, Barman, 2007), CO, CO, HO and CH (Stevenson et al., 2010). Swain et al. (2009) also measure signatures of CH, CO and HO using NICMOS.
Secondary eclipse observations also place constraints on other physical parameters such as atmospheric temperatures and reflectivity. Knutson et al. (2008) interpreted thermal emission measured with IRAC at 3.6, 4.5, 5.8 and 8.0 microns as requiring a thermal inversion layer high in the atmosphere, as well as confirming the detection of water. This indicates that there must be absorbing species in this region of the atmosphere, which trap stellar heat and cause the atmosphere to heat up. Burrows et al. (2007) compared their models to the secondary eclipse data, and also found that they required the day-side atmosphere to have a stratospheric temperature inversion. A consequence of this heating is molecular emission bands from the planet, which explains the water detection in emission. Burrows et al. (2008) further confirmed the requirement for the planet to have an inversion. They also showed that this is not the case for all planets, and showed that there are two different populations, those with inversions and those without. HD 189733b is an example of a planet without a stratospheric inversion. Whilst the low emission features from previous observations were hard to match to previous models, Burrows et al. (2008) showed that models with an extra upper atmosphere absorber in the optical are consistent with the data, particularly at high values of Pn (high efficiency of energy transportation from dayside to nightside).
Optical measurements in secondary eclipse are sensitive to reflected starlight and can hence constrain the optical reflective properties of planets. As with previous secondary eclipse observations, these measurements are very difficult. Snellen (2005) found no firm detection of light from the planet in the K band. Further non-detections have provided upper limits for what seems to be a very low optical albedo. Rowe et al. (2006, 2008) provided constraints using the MOST satellite. Their non-detection places a 3 sigma upper limit of 0.17 on the albedo. The measured upper limit is significantly less reflective than Jupiter (Ag ~ 0.5) and rules out the presence of bright reflective clouds in the planet’s atmosphere. Burrows et al. (2008) compared albedo models for close-in giant planets with this upper limit and found that all models without scattering can fit the observations. This finding is also consistent with the sodium line profile, which shows the presence of line wings, indicating a relatively clear upper atmosphere.
Observations of changes in the planetary brightness as it orbits its star can give more direct information about the efficiency of heat redistribution. Cowan et al. (2007) attempt to measure mid infrared (3.6 or 4.5 and 8 micron) phase variations of several hot Jupiters. Due to significant systematic errors between different epochs, they can only place a 2 sigma upper limit of 0.0015 for HD 209458 b (compared to 0.0007 for 51 Peg). However, this observation still gives the constraint that HD 209458 b must recirculate at least 32% of incident stellar light to its night side at the 1 sigma level.
Another exciting observation from transmission spectroscopy of HD 209458 b is the detection of the first escaping atmosphere. Vidal-Madjar et al. (2003) detected an absorption depth in Lyman- of 15%, which indicates atmospheric material beyond the Roche lobe of the planet. They estimate a mass-loss rate g/s. The lines are also blueshifted, indicating that material is being blown away from the direction of the host star (Ballester et al., 2007). Evaporation rates from models (Lecavelier des Etangs et al., 2004, 2007; Schneiter et al., 2007 Murry-Clay et al., 2009) are consistent with the observational measurements. Ehrenreich et al. (2008) measure the transit depth in Lyman- using ACS and also find consistent results. H I, O I, C II, Si III and marginally Si IV are also detected with large signatures, indicating that these species are also present in the escaping atmosphere (Vidal-Madjar et al., 2004; Linsky et al., 2010, Schlawin et al., 2010). The detection of a large escaping atmosphere prompted investigations of upper atmospheric chemistry. Liang et al. (2003) find that the production of atomic hydrogen is mainly driven by HO photolysis and the reaction of OH with H and insensitive to exact abundances of CO, HO, CH.
This result is still debated. Ben-Jaffel (2007) found only an 8.9% absorption depth using HST data, and no evidence of atmospheric escape. Vidal-Madjar et al. (2008) pointed out that the two measurements were obtained at different wavelengths and are indeed consistent once this is taken into account. Ben-Jaffel (2008), however, show that absorption depth during transit is not sensitive to the assumed spectral range in the stellar line profile, and confirm their previous result. They also find from the line profile that the inflated hydrogen atmosphere around the planet is symmetric and typical of a Lorentzian, optically thick medium. They find no asymmetry between the blue and red sides of the transit. Hölmstrom et al. (2008) suggest another explanation for the large spectral signature. Solar system planets show Lyman- absorption due to interaction of the stellar wind with the exosphere rather than escape of the planetary atmosphere, and this could also be the case for HD 209458 b.
HD 209458 b belongs to the class of bloated hot Jupiters, of which there are now many discovered. Particular observations for HD 209458 b have ruled out some contributions to tidal heating. Observations have now progressively placed limits on potential planetary companions to be less than one Earth mass (Brown et al., 2001, Croll et al., 2007, Steffen & Agol et al., 2007, Miller-Ricci et al., 2008). The eccentricity and misalignment are independently measured to be small (Bodenheimer et al., 2001, Laughlin et al., 2005, Fabrycky et al., 2007). It is, however, still possible that tidal heating in the history of the planet could have played a role (Jackson et al., 2008, Guo et al., 2010).
Our own giant planets have measurable magnetic fields, and it has recently begun to be investigated whether HD 209458 b also has one. However, current observations have resulted in non-detections of the auroral emission at 1100-1700 Angstroms and emission at 150 MHz, indicating a weak magnetic field (France et al., 2010; Lecavelier des Etangs et al., 2011).