HAT-P-55b: A Hot Jupiter Transiting a Sun-Like Star¹

Today we know of almost 1800 validated exoplanets and more than 4000 exoplanet candidates. Among these, the transiting exoplanets (TEPs) are essential in our exploration and understanding of the physical properties of exoplanets. While radial velocity observations alone only allow us to estimate the minimum mass of a planet, we can combine them with transit observations for a more comprehensive study of the physical properties of the planet. When a planet transits, we can determine the inclination of the orbit and the radius of the planet, allowing us to break the mass degeneracy and, along with the mass, determine the mean density of the planet. The mean density of a planet offers us insight into its interior composition, and although there is an inherent degeneracy arising from the fact that planets of different compositions can have identical masses and radii, this information allows us to map the diversity and distribution of exoplanets and even put constraints on models of planetary structure and formation theories.

The occurrence rate of hot Jupiters in the Solar neighborhood is around 1% (Udry & Santos 2007; Wright et al. 2012). With a transit probability of about ∼10%, roughly one thousand stars need to be monitored in photometry to find just a single hot Jupiter. Therefore, the majority of the known transiting hot Jupiters have been discovered by photometric wide field surveys targeting tens of thousands of stars per night.

In this paper, we present the discovery of HAT-P-55b by the Hungarian-made Automated Telescope Network (HATNet; see Bakos et al. [2004]), a network of six small fully automated wide field telescopes of which four are located at the Fred Lawrence Whipple Observatory in Arizona, and two are located at the Mauna Kea Observatory in Hawaii. Since HATNet saw first light in 2003, it has searched for TEPs around bright stars (V ≲ 13) covering about 37% of the Northern sky, and discovered approximately 25% of the known transiting hot Jupiters.

The layout of the paper is as follows. In § 2, we present the different photometric and spectroscopic observations that lead to the detection and characterization of HAT-P-55b. In § 3, we derive the stellar and planetary parameters. Finally, we discuss the characteristics of HAT-P-55b in § 4.

The general observational procedure used by HATNet to discover TEPs has been described in detail in previous papers (e.g., Bakos et al. 2010; Latham et al. 2009). In this section, we present the specific details for the discovery and follow-up observations of HAT-P-55b.

2.1. Photometry

2.1.1. Photometric Detection

HAT-P-55b was initially identified as a candidate transiting exoplanet based on photometric observations made by the HATNet survey (Bakos et al. 2004) in 2011. The observations of HAT-P-55 were made on nights between February and August with the HAT-5 telescope at the Fred Lawrence Whipple Observatory (FLWO) in Arizona, and on nights between May and August with the HAT-8 telescope at Mauna Kea Observatory in Hawaii. Both telescopes used a Sloan r-band filter. HAT-5 provided a total of 10,574 images with a median cadence of 218 s, and HAT-8 provided a total of 6428 images with a median cadence of 216 s.

The results were processed and reduced to trend-filtered light curves using the External Parameter Decorrelation method (EPD; see Bakos et al. 2010) and the Trend Filtering Algorithm (TFA; see Kovács et al. 2005). The light curves were searched for periodic transit signals using the Box Least-Squares method (BLS; see Kovács et al. 2002). The individual photometric measurements for HAT-P-55 are listed in Table 1, and the folded light curves together with the best-fit transit light curve model are presented in Figure 1.

Zoom In Zoom Out Reset image size

The full table is available.

2.1.2. Photometric Follow-Up

We performed photometric follow-up observations of HAT-P-55 using the KeplerCam CCD camera on the 1.2-m telescope at the FLWO, observing a transit ingress on the night of 2013 May 23, and a full transit on the night of 2014 April 7. Both transits were observed using a Sloan i-band filter. For the first event we obtained 230 images with a median cadence of 64 s, and for the second event we obtained 258 images with a median cadence of 67 s.

The results were reduced to light curves following the procedure of Bakos et al. (2010), and EPD and TFA were performed to remove trends simultaneously with the light curve modeling. The individual photometric follow-up measurements for HAT-P-55 are listed in Table 1, and the folded light curves together with our best-fit transit light curve model are presented in Figure 2.

Zoom In Zoom Out Reset image size

Subtracting the transit signal from the HATNet light curve, we used the BLS method to search for additional transit signals and found none. A Discrete Fourier Transform also revealed no other periodic signals in the data.

2.2. Spectroscopy

We performed spectroscopic follow-up observations of HAT-P-55 to rule out false positives and to determine the RV variations and stellar parameters. Initial reconnaissance observations were carried out with the Tillinghast Reflector Echelle Spectrograph (TRES; Fűrész 2008) at the FLWO. We obtained two spectra near opposite quadratures on the nights of October 4 and 31, 2012. Using the Stellar Parameters Classification method (SPC; see Buchhave et al. [2012]), we determined the initial RV measurements and stellar parameters. We found a mean absolute RV of -9.42 km s^-1 with an rms of 48 ms^-1, which is consistent with no detectable RV variation. The stellar parameters, including the effective temperature T_eff⋆ = 5800 ± 50 K, surface gravity log g_⋆ = 4.5 ± 0.1 (log cgs), and projected rotational velocity v sin i = 5.0 ± 0.4 km s^-1, correspond to those of a G2 dwarf.

High-resolution spectroscopic observations were then carried out with the SOPHIE spectrograph mounted on the 1.93-m telescope at Observatoire de Haute-Provence (OHP) (Perruchot et al. 2011; Bouchy et al. 2013), and with the FIES spectrograph mounted on the 2.6-m Nordic Optical Telescope (Djupvik & Andersen 2010). We obtained six SOPHIE spectra on nights between 2013 June 3 and 12, and ten FIES spectra on nights between May 15, 2013 and August 26, 2013.

We reduced and extracted the spectra and derived radial velocities and spectral line bisector span (BS) measurements following the method of Boisse et al. (2013) for the SOPHIE data and the method of Buchhave et al. (2010) for the FIES data. The final RV data and their errors are listed for both instruments in Table 2, and the folded RV data together with our best-fit orbit and corresponding residuals and bisectors are presented in Figure 3. To avoid underestimating the BS uncertainties we base them directly on the RV uncertainties, setting them equal to twice the RV uncertainties. At a first glance, there does seem to be a slight hint of variation of the BSs in phase with the RVs suggesting there might be a blend. This is not the case, however, as will we show in our detailed blend analysis in § 3.

Zoom In Zoom Out Reset image size

We applied the SPC method to the FIES spectra to determine the final spectroscopic parameters of HAT-P-55. The values were calculated using a weighted mean, taking into account the cross correlation function (CCF) peak height. The results are shown in Table 3.

In order to rule out the possibility that HAT-P-55 is a blended stellar eclipsing binary system, and not a transiting planet system, we carried out a blend analysis following Hartman et al. (2012). We find that a single star with a transiting planet fits the light curves and catalog photometry better than models involving a stellar eclipsing binary blended with light from a third star. While it is possible to marginally fit the photometry using a G star eclipsed by a late M dwarf that is blended with another bright G star, simulated spectra for this scenario are obviously composite and show large (multiple km s^-1) bisector span and RV variations that are inconsistent with the observations. Based on this analysis we conclude that HAT-P-55 is not a blended stellar eclipsing binary system, and is instead best explained as a transiting planet system. We also consider the possibility that HAT-P-55 is a planetary system with a low-mass stellar companion that has not been spatially resolved. The constraint on this scenario comes from the catalog photometric measurements, based on which we can exclude a physical companion star with a mass greater than 0.7 M_⊙. Any companion would dilute both the photometric transit and radial velocity orbit. The maximum dilution allowed by the photometry would increase the planetary radius by ∼15%.

We analyzed the system following the procedure of Bakos et al. (2010) with modifications described in Hartman et al. (2012). In short, we (1) determined the stellar atmospheric parameters of the host star HAT-P-55 by applying the SPC method to the FIES spectra; (2) used a Differential Evolution Markov-Chain Monte Carlo procedure to simultaneously model the RVs and the light curves, keeping the limb-darkening coefficients fixed to those of the Claret (2004) tabulations; and (3) used the spectroscopically inferred effective temperatures and metallicities of the star, the stellar densities determined from the light curve modeling, and the Yonsei-Yale theoretical stellar evolution models (Yi et al. 2001) to determine the stellar mass, radius, and age, as well as the planetary parameters (e.g., mass and radius) which depend on the stellar values (Fig. 4).

Zoom In Zoom Out Reset image size

We conducted the analysis inflating the SOPHIE and FIES RV uncertainties by adding a "jitter" term in quadrature to the formal uncertainties. This was done to accommodate the larger-than-expected scattering of the RV observations around the best-fit model. Independent jitters were used for each instrument, as it is not clear whether the jitter is instrumental or astrophysical in origin. The jitter term was allowed to vary in the fit, yielding a χ² per degree of freedom of unity for the RVs in the best-fit model. The median values for the jitter are 26 ± 13 ms^-1 for the SOPHIE observations and 0.0 ± 3.1 ms^-1 for the FIES observations. This suggests that either the formal uncertainties of the FIES instrument were overestimated, or that the jitter from the SOPHIE instrument is not from the star but from the instrument itself.

The analysis was done twice: fixing the eccentricity to zero, and allowing it to vary. Computing the Bayesian evidence for each model, we found that the fixed circular model is preferred by a factor of ∼500. Therefore the circular orbit model was adopted. The 95% confidence upper limit on the eccentricity is e < 0.139.

The best-fit models are presented in Figures 1, 2 and 3, and the resulting derived stellar and planetary parameters are listed in Tables 3 and 4, respectively. We find that the star HAT-P-55 has a mass of 1.013 ± 0.037 M_⊙ and a radius of 1.011 ± 0.036 R_⊙, and that its planet HAT-P-55b has a period of 3.5852467 ± 0.0000064 days, a mass of 0.582 ± 0.056 M_J, and a radius of 1.182 ± 0.055 R_J.

We have presented the discovery of a new transiting planet, HAT-P-55b, and provided a precise characterization of its properties. HAT-P-55b is a moderately inflated ∼0.5 M_J planet, similar in mass, radius, and equilibrium temperature to HAT-P-1b (Bakos et al. 2007), WASP-34 (Smalley et al. 2011), and HAT-P-25b (Quinn et al. 2012).

With a visual magnitude of V = 13.21, HAT-P-55b is among the faintest transiting planet host stars discovered by a wide field ground-based transit survey (today, a total of 11 transiting planet host stars with V > 13 have been discovered by wide field ground-based transit surveys, the faintest one is HATS-6 with V = 15.2 [Hartman et al. 2014]). Of course, V > 13 is only faint by the standards of surveys like HATNet and WASP; most of the hundreds of transiting planets found by space-based surveys such as OGLE, CoRoT, and Kepler have host stars fainter than HAT-P-55. It is worth noticing that despite the relative faintness of HAT-P-55, the mass and radius of HAT-P-55b has been measured to better than 10% precision (relative to the precision of the stellar parameters) using modest-aperture facilities. This achievement was possible because the relatively large size of the planet and its close proximity to its host star provided for a strong and therefore easy-to-measure signal. In comparison, only about 140 of all the 1175 known TEP's have masses and radii measured to better than 10% precision.

• Table S1