Observations

In this section, we summarize the important spectroscopic and photometric observations of transiting planets that have been conducted, during both primary and secondary eclipse. Most of these observations have been performed on the planet HD 209458 b, since it was detected first. We conclude by describing how the model calculations have helped to interpret these results.

As discussed in Section 2.2, the planetary spectrum can be probed during transit using a method called transmission spectroscopy. Although the planetary spectrum is ~ 10 000 times fainter than that of the star, the differential nature of the measurement makes it possible to achieve this precision. Several detections and useful upper limits have been obtained on HD 209458 b:

• sodium doublet detected (Charbonneau et al., 2002);

• hydrogen Lyman-a detected (Vidal-Madjar et al., 2003);

• carbon monoxide upper limit (Deming et al., 2005a).

The amount of sodium detected was approximately a factor of 3 smaller than expected from simple models of the atmosphere, suggesting the presence of a high cloud that masks the true sodium abundance. The detection of the transit in H Lyman-a was very significant - a 15% drop in stellar flux during transit, 10 times greater than the transit depth at visible wavelengths. This implies an extended atmosphere of 3 or 4 Jupiter radii, and suggests that the planet is losing mass over its lifetime. The CO non-detection further reinforces the notion of a high cloud in the planet's atmosphere.

2.7 Observations

The complementary technique during secondary eclipse is called occultation spectroscopy. Briefly, this involves taking spectra of the system when the planet is out of eclipse (when both the star and planet are visible) and comparing these with spectra recorded when the planet is hidden during secondary eclipse. By carefully differencing these spectra, one can, in principle, derive the spectrum of the planet itself. Although this technique has not yet been successfully conducted on extrasolar planets, early attempts have yielded some useful information:

• upper limit on emission near 2.2 |m (Richardson et al., 2003b);

• upper limit on methane abundance (Richardson et al., 2003a).

Both of these limits were derived from ground-based observations, which are often limited by variations in the terrestrial atmosphere, making detection of spectral features difficult.

We now turn to photometric observations of the secondary eclipse that have occurred most recently. Although measurable, the effect due to the secondary eclipse is small, e.g., ~0.3% for HD 209458 b at 20 |m (see Figure 2.2), and decreasing for smaller wavelengths. NASA's Spitzer Space Telescope2 was responsible for the first detection of a secondary eclipse of a transiting planet. Spitzer, with an 85-cm aperture, has three instruments on board that together perform photometry and spectroscopy at infrared wavelengths. In March 2005, two independent research groups announced detections of the secondary eclipse of two different planets using two Spitzer instruments. Observations of HD 209458 b with the Multiband Imaging Photometer for Spitzer (MIPS) detected the secondary eclipse at 24 |m (Deming et al., 2005b), while TrES-1 was observed in two wavelengths (4.5 and 8 |m) with the Infrared Array Camera (IRAC) (Charbonneau etal., 2005). These observations represent the first direct detection of an extrasolar planet. More recently, the secondary eclipse of HD 189733 b was observed at 16 |m using the Infrared Spectrograph (IRS), although the observation was performed photometrically, not spectroscopically, using a detector that is normally used only to align the star on the slit (Deming et al., 2006).

The secondary eclipse detections provide a measurement of the brightness temperature of the planets, at the respective wavelengths. The brightness temperature is the blackbody temperature of an object at a particular wavelength; given the irradiance, the blackbody function (see Equation (2.6)) can be inverted to solve for temperature. For HD 209458 b the brightness temperature at 24 |m is 1130 ± 150 K (Deming etal., 2005b), and for TrES-1 it is 1010 ± 60 K at 4.5 |im and 1230 ± 110 K at 8 |im (Charbonneau et al., 2005). HD 189733 b has a brightness temperature of 1117 ±42 K at 16 |m (Deming et al., 2006). Although models

2 http://ssc.spitzer.caltech.edu/

have predicted the effective temperature of the atmospheres of extrasolar planets, these are the first observational measurements of the temperature of an extrasolar planetary atmosphere.

With the Spitzer photometry of several transiting planets, as well as ground-based spectroscopic observations, we can now compare the observational results to theoretical calculations and begin to construct a comprehensive picture of the atmospheres of the transiting planets. In the wake of the three initial photometric detections of thermal emission from two extrasolar planets (Deming et al., 2005b; Charbonneau et al., 2005), four theory papers (Seager et al., 2005; Barman et al., 2005; Fortney et al., 2005; Burrows et al., 2005) appeared within a few months to explain the results! Some of these even have conflicting conclusions. One conclusion on which all of the explanations agree is that the planets are hot, as predicted. (We note that this was not a given; a planet with a high Bond albedo, for example, would reflect much of the incident stellar irradiation and therefore could be much cooler, as seen in Equation (2.19).)

The second point on which all modellers agree is that the TrES-1 data points at 4.5 and 8.0 ^m are not consistent with the assumption of solar abundances, because the 8.0 ^m flux is too high. Beyond these two conclusions, the interpretations diverge.

Seager et al. (2005) conclude that a range of models remain consistent with the data. They include the 2.2 ^m upper limit reported by Richardson et al. (2003b) (which has been largely ignored by modellers), as well as an upper limit on the albedo from the Canadian MOST satellite (Rowe et al., 2005), and are able to eliminate the models for HD 209458 b that are on the hot and cold ends of the plausible temperature range for the planet. Their work suggests that an intermediate value for f (see Equation (2.19)) is most likely, indicating that the atmospheric circulation is somewhere between the two extremes (efficient redistribution vs none at all). The interpretation by Seager et al. (2005) and the observational results for HD 209458 b are shown in Figure 2.4.

Fortney et al. (2005) show that standard models using solar abundances are consistent with HD 209458 b but only marginally consistent (within 2a at 8.0 ^m) for TrES-1. For both planets, their best fit models assume that the incident stellar radiation is redistributed efficiently throughout the atmosphere (i.e., f = 1). On the other hand, Burrows et al. (2005) conclude that the f = 2 case is more likely, indicating that the day side is significantly brighter in the infrared than the night side. They also infer the presence of CO and possibly H2O. The resolution of this discrepancy awaits further Spitzer observations.

Finally, with the recent Spitzer detection of HD 189733 b during secondary eclipse at 16 ^m (Deming et al., 2006) and detections by IRAC and MIPS under analysis (D. Charbonneau, private communication), we have more data available for comparison with theoretical spectra. In addition, observations are being analysed

2.8 Future missions

2.8 Future missions

Wavelength (|m)

Wavelength (|m)

Fig. 2.4. Theoretical spectrum of HD 209458 b with data points and upper limits. The solid curve is a cloud-free model with solar abundances and f = 2, characterized by deep water vapour absorption features. From left to right the data are: the MOST upper limit (Rowe et al., 2005), a constraint on the H2O band depth (Richardson et al., 2003a; Seager et al., 2005), and the Spitzer/MIPS thermal emission point at 24 ^m (Deming et al., 2005b). The solid lines show 1 a error bars or upper limits and the dashed lines show 3 a values. Note the linear flux scale on the upper panel and the log flux scale on the lower panel.

Wavelength (|m)

Fig. 2.4. Theoretical spectrum of HD 209458 b with data points and upper limits. The solid curve is a cloud-free model with solar abundances and f = 2, characterized by deep water vapour absorption features. From left to right the data are: the MOST upper limit (Rowe et al., 2005), a constraint on the H2O band depth (Richardson et al., 2003a; Seager et al., 2005), and the Spitzer/MIPS thermal emission point at 24 ^m (Deming et al., 2005b). The solid lines show 1 a error bars or upper limits and the dashed lines show 3 a values. Note the linear flux scale on the upper panel and the log flux scale on the lower panel.

or planned to detect a mid-infrared emission spectrum of HD 209485 b and HD 189733 b, respectively, both using Spitzer/IRS. These observations would be the first observed emission spectra of an extrasolar planet and will advance our understanding beyond the few photometric data points we have now.

Was this article helpful?

0 0

Post a comment