WASP-35b, WASP-48b, AND HAT-P-30b/WASP-51b: TWO NEW PLANETS AND AN INDEPENDENT DISCOVERY OF A HAT PLANET

The WASP-North and South observatories are located at La Palma, Canary Islands, and SAAO, South Africa, respectively, and each consist of eight 11 cm telescopes of 78 × 78 field of view each, on a single fork mount. The cameras scan repeatedly through eight to ten sets of fields each night, taking 30 s exposures. See Pollacco et al. (2006) for more details on the WASP project and the data reduction procedure, and Collier Cameron et al. (2007) and Pollacco et al. (2008) for an explanation of the candidate selection process.

High-precision transit light curves of WASP-35, WASP-48, and HAT-P-30/WASP-51 were obtained with the Rapid Imager to Search for Exoplanets (RISE) instrument mounted in the 2.0 m Liverpool Telescope (Steele et al. 2008; Gibson et al. 2008) at the Observatorio del Roque de los Muchachos on La Palma, Canary Islands. The data were reduced with the ULTRACAM pipeline (Dhillon et al. 2007) following the same procedure as Barros et al. (2011). Further high-quality photometry of WASP-35 was obtained with the 0.6 m TRAnsiting Planets and PlanetesImals Small Telescope (TRAPPIST) at La Silla Observatory, Chile, and with the LCOGT 2.0 m Faulkes Telescope South (FTS) at Siding Spring Observatory.

Photometric observations for each of WASP-35, WASP-48, and HAT-P-30/WASP-51 are detailed in Table 2 and shown in Figures 1–3.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We obtained three radial velocity (RV) measurements of WASP-35 on 2010 January 25 and 26 with the FIbre-fed Echelle Spectrograph (FIES) on the 2.5 m Nordic Optical Telescope (NOT).¹⁵ FIES was used in medium resolution mode (R = 46,000) with simultaneous ThAr calibration, and the observations were reduced using the FIEStool package,¹⁶ then cross-correlated with a high signal-to-noise spectrum of the Sun to obtain the radial velocities. A further nine data points were obtained with the CORALIE echelle spectrograph mounted on the Swiss telescope at the ESO La Silla Observatory in Chile on 2010 January 5 and between 2010 February 6 and 14 and the resulting spectra were processed with the standard data reduction CORALIE pipeline (Baranne et al. 1996; Mayor et al. 2009), plus a correction for the blaze function (Triaud 2010).

WASP-48 was observed spectroscopically with the SOPHIE fiber-fed echelle spectrograph mounted on the 1.93 m telescope at Observatoire de Haute-Provence (OHP) (Perruchot et al. 2008; Bouchy et al. 2009). Fourteen data points were taken in high efficiency mode (R = 40,000) between 2010 April 19 and October 18, obtaining a signal-to-noise ratio (S/N) of around 35 for each. Wavelength calibration with a thorium–argon lamp was performed to allow the interpolation of the spectral drift of SOPHIE.

Radial velocity observations of HAT-P-30/WASP-51 were made with the CORALIE and SOPHIE spectrographs. Nine data points were taken with SOPHIE between 2010 October 16 and November 25, using the high-resolution mode (R = 75,000) and obtaining an S/N of around 20 for each measurement. Ten further data points were obtained with CORALIE between 2010 December 7 and 2011 January 4.

The radial velocities measured using these spectra are given in Tables 3–5 and shown as a function of orbital phase in Figures 4–6, along with radial velocity residuals and bisector span variations. The radial velocity measurements show low-amplitude variations, compatible with the existence of orbiting planets, and are not correlated with the bisector spans (Queloz et al. 2001), available for CORALIE and SOPHIE data (no correlation is found with p-value less than 0.05), which could indicate stellar blend configurations.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The CORALIE guiding camera revealed a faint star located at about 15 from HAT-P-30/WASP-51 which was observed simultaneously with HAT-P-30/WASP-51 through the 3'' fiber of the SOPHIE spectrograph. This target was too close and too faint to perform aperture photometry with the CORALIE guiding images in order to estimate the magnitude. Thus, we performed radial velocity blend simulations (A. Santerne et al. 2011, in preparation) in order to investigate whether small bisector variations seen in SOPHIE data are due to this secondary blending star, though the bisector spans are compatible with no variation within the 2σ level. This simulation consisted of producing SOPHIE cross-correlation functions (CCFs) for HAT-P-30/WASP-51 with the unblended orbital solution seen by CORALIE and blending a single (solar) star with different systemic radial velocity, distance ratio from HAT-P-30/WASP-51, and vsin i. Since the distance between the two stars is degenerate with their surface brightness different, and without any magnitude constraint on the second star, we assumed it to be a solar-type star in order to simplify the simulation. We carry out 10,000 Monte Carlo simulations with the following input parameters: the systemic RV was chosen in a range 34.6–54.6 km s⁻¹ since the two stars have to be well aligned in velocity to reproduce such a symmetric bisector effect, the distance ratio of HAT-P-30/WASP-51 was chosen between 1 and 5 (a solar star five times further from HAT-P-30/WASP-51 is too faint to significantly affect its CCF), and finally we test the vsin i of the blending star between 4 and 40 km s⁻¹. We did not find any significant model that reproduces the bisector effect seen in the SOPHIE data with the observed bisector amplitude. To reproduce such a bisector span, it is necessary to have much larger semi-amplitude RV variations coupled with an RV-aligned blend which is not compatible with the CORALIE nor HIRES (Johnson et al. 2011) data. In the case of HAT-P-30/WASP-51, the effect of the blending star on the bisector is at a level of a few m s⁻¹. We concluded that the faint star 15 away did not significantly affect the SOPHIE CCF of HAT-P-30/WASP-51.

The individual NOT/FIES spectra of WASP-35 were co-added to produce a single spectrum with an average S/N of around 100:1. For WASP-48, the SOPHIE spectra were co-added to produce a spectrum with a typical S/N of around 80:1. For HAT-P-30/WASP-51, the CORALIE spectra were co-added to produce a single spectrum with an average S/N of around 150:1. The standard pipeline reduction products were used in the analysis.

The analysis was performed using the methods given in Gillon et al. (2009). The Hα line was used to determine the effective temperature, T_eff, while the Na i D and Mg i b lines were used as surface gravity, log g  diagnostics. The elemental abundances were determined from equivalent-width measurements of several clean and unblended lines. Atomic line data were mainly taken from the Kurucz & Bell (1995) compilation, but with updated van der Waals broadening coefficients for lines in Barklem et al. (2000) and log gf values from Gonzalez & Laws (2000), Gonzalez et al. (2001), and Santos et al. (2004). A value for microturbulence, ξ_t, was determined from Fe i using the method of Magain (1984). The parameters obtained from the analyses are listed in Table 6. The quoted error estimates include those given by the uncertainties in T_eff, log g, and ξ_t, as well as the scatter due to measurement and atomic data uncertainties.

WASP-51b has recently been announced as HAT-P-30b (Johnson et al. 2011). A discrepancy between their results and ours is the value of host star metallicity: we find a value of [Fe/H] = −0.08 ± 0.08, whereas Johnson et al. (2011) report [Fe/H] = 0.13 ± 0.08, from an analysis using Spectroscopy Made Easy (SME; Valenti & Piskunov 1996). The difference is partly due to a different solar metallicity value adopted in the spectral analysis, where we use log A(Fe) = 7.54 ± 0.03 from Biemont et al. (1991), which is 0.04 dex higher than the current Asplund et al. (2009) solar value of log A(Fe) = 7.50 ± 0.04. We continue to use the Biemont et al. (1991) value for consistency between analyses rather than adjusting to new values as they are announced. Additionally, we allow the microturbulence value to float and be fitted, finding a best-fit value of 1.3 km s⁻¹, in excellent agreement with the calibration in Bruntt et al. (2010), compared to the assumed value of 0.85 km s⁻¹ adopted by Valenti & Fischer (2005) and used in Johnson et al. (2011). Using the lower value of microturbulence would increase our [Fe/H] abundance by 0.07. The slight differences in T_eff and log g values, where Johnson et al. (2011) find T_eff = 6300 ± 90 and log g = 4.36 ± 0.03, result in a difference of a further 0.03 dex. These three factors combine to reduce the discrepancy to only 0.07 dex, which is within the quoted uncertainties.

The projected stellar rotation velocity (vsin i) was determined by fitting the profiles of several unblended Fe i lines. Values for macroturbulence, v_mac, of 3.3 ± 0.3 km s⁻¹, 3.2 ± 0.3 km s⁻¹, and 4.0 ± 0.3 km s⁻¹ were assumed for WASP-35, WASP-48, and HAT-P-30/WASP-51, respectively, based on the calibration by Bruntt et al. (2010), and instrumental FWHM of 0.13 ± 0.01 Å, 0.15 ± 0.01 Å, and 0.11 ± 0.01 Å, respectively, determined from the telluric lines around 6300 Å. Best-fitting values of vsin i = 3.9 ± 0.4 km s⁻¹, 12.2 ± 0.7 km s⁻¹, and 3.6 ± 0.4 km s⁻¹, respectively, were obtained. This value for vsin i for HAT-P-30b/WASP-51b agrees with the value of 3.1 ± 0.2 km s⁻¹ found by Johnson et al. (2011) via their fitting of the Rossiter–McLaughlin effect for this system, while they found a lower value of 2.2 ± 0.5 km s⁻¹ from spectral line fitting. An alternative determination of the macroturbulence based on the tabulation by Gray (2008), rather than the new calibration from Bruntt et al. (2010), results in v_mac = 5.1 ± 0.3 km s⁻¹, leading to vsin i = 1.8 ± 0.4 km s⁻¹, which agrees with the value found by Johnson et al. (2011).

The lithium abundance in WASP-35 implies an age of ≳2 Gyr according to the calibration of Sestito & Randich (2005), while the measured vsin i gives a rotational period of P_rot ≃ 14 ± 2 days, assuming i = 90°, which yields gyrochronological age of ∼2.2^+1.4_−0.9 Gyr using the relation of Barnes (2007). A lack of stellar activity is indicated by the absence of Ca ii H+K emission in the spectra.

The non-detection of lithium in the WASP-48 spectrum suggests that the star is several gigayears old, and the lack of any Ca H+K emission is consistent with this. However, the stellar rotation rate of 4.5 ± 0.6 days from the vsin i measurement implies an age of only ∼0.2^+0.2_−0.1 Gyr using the Barnes (2007) gyrochronological relation. This discrepancy is discussed further below.

The presence of strong lithium absorption in the HAT-P-30/WASP-51 spectrum suggests that the star is ≲1 Gyr old. The stellar rotation period of 18.0 ± 3.1 days found from the vsin i measurement results in an age of ∼5.6^+7.2_−2.9 Gyr, although this can be regarded as an upper limit, since the stellar rotation rate may be higher due to the unknown value of stellar inclination (i).

We simultaneously analyzed all photometry and radial velocity data using a Markov Chain Monte Carlo (MCMC) analysis, as set out in Collier Cameron et al. (2007), modified to calculate the mass of the host star with a calibration on T_eff, log ρ, and [Fe/H] as described in Enoch et al. (2010). The photometry provides the stellar density (assuming the eccentricity of the planetary orbit is known) while the effective temperature and metallicity of the star are obtained through spectral analysis given above.

We initially allowed the value of the eccentricity of each planetary orbit to float, which resulted in values of e = 0.057^+0.064_−0.032, e = 0.058^+0.058_−0.035, and e = 0.040^+0.044_−0.028 for WASP-35b, WASP-48b, and HAT-P-30b/WASP-51b, respectively. However, in each case the χ² values in fitting the radial velocity data are lower for a circular orbit than the eccentric solutions found here: the eccentric solutions produce radial velocity χ² values of 4.5, 15.4, and 10.6 for WASP-35, WASP-48, and HAT-P-30/WASP-51, respectively, for 12, 14, and 20 RV data points, while a circular orbit for each results in χ² values of 0.60, 8.6, and 1.8. With no strong evidence for eccentricity in the orbits, we then performed second analyses of all systems, fixing the eccentricities to zero.

The MCMC results for WASP-48b reveal a slightly evolved star with a large radius of 1.75 R_☉ (discussed below). This radius is around 0.6 R_☉ greater than expected for that of a star of the same mass on the main sequence. To explore this, we performed a further analysis, imposing a main-sequence constraint on WASP-48, such that a Gaussian prior is used on the stellar radius at each step in the analysis, centered on radius R_* = M^0.8_*, again allowing the eccentricity to float. This forced the stellar radius estimate down to 1.13 R_☉, artificially inflating the eccentricity estimate to 0.38 ± 0.04 to compensate, since the stellar radius is calculated from the photometry, including a term $\frac{1 + e \sin \omega }{\sqrt{1-e^2}}$: an eccentricity of 0.4 produces a stellar radius value 0.65 times that of the estimate with eccentricity at zero. However, there is no evidence for such a high eccentricity in the radial velocity curve (see Figure 5) and the χ² values are 33.3 and 8.6 for eccentric and circular orbits, respectively (where there are effectively four degrees of freedom: radial velocity semi-amplitude, K, center-of-mass velocity, γ, eccentricity, e, and argument of periastron, ω). Imposing the main-sequence constraint with the eccentricity held fixed at zero decreases the stellar radius to 1.31 ± 0.09 R_☉ and decreases the impact parameter from 0.7 to 0.4 to compensate: such a reduction in impact parameter leads directly to a stellar radius estimate 0.70 times smaller. However, the much lower impact parameter estimate produces a model that does not match well the high-quality follow-up photometric data: compare the fit to ingress and egress in Figure 7, showing the lower impact parameter model, with that in Figure 2. Since neither analysis with main-sequence constraint imposed produces realistic models for the photometry and spectroscopy, the results from the analysis with no main-sequence constraint are clearly preferred.

Zoom In Zoom Out Reset image size

We also tested the imposition of the main-sequence constraint on WASP-35 and HAT-P-30/WASP-51, producing negligible change in all parameters, well within the parameter uncertainties, which implies that the photometry is sufficient to constrain well the results.

The best-fit parameters for each system, with eccentricity fixed to zero and no main-sequence constraint imposed, are given in Table 7, where HJD values are in HJD_UTC—see Eastman et al. (2010) for conversion to alternative systems—and uncertainty values are from analyses where the follow-up photometry is binned to one-third of the ingress/egress duration (8, 10, and 14 minute bins for WASP-35, WASP-48, and WASP-51, respectively) to account for correlated noise (longer bins result in too few points in ingress/egress and hence large uncertainty in impact parameter). WASP-35 is found to have a mass of 1.07 ± 0.03 M_☉ and a radius of 1.09 ± 0.03 R_☉, while WASP-35b has a mass of 0.72 ± 0.06 M_J and a radius of 1.32 ± 0.05 R_J, giving a density of 0.32 ± 0.04ρ_J. For WASP-48, the best-fit model gives a stellar mass of 1.19 ± 0.05 M_☉ and a radius of 1.75 ± 0.09 R_☉. WASP-48b is found to have a mass of 0.98 ± 0.09 M_J and a radius of 1.67 ± 0.10 R_J, giving a density of 0.21 ± 0.04ρ_J.

The best-fit result for HAT-P-30/WASP-51 gave a stellar mass 1.18 ± 0.03 M_☉, stellar radius 1.33 ± 0.03 R_☉, planetary mass 0.76 ± 0.05 M_J, and planetary radius 1.42 ± 0.03 R_J, giving a density of 0.26 ± 0.03ρ_J. Under the name HAT-P-30b, based on HAT-5 (Arizona) and HAT-9 (Hawaii) photometry of a total of around 3,200 good data points, follow-up high-quality photometric observations with Kepler-Cam on the FLWO 1.2 m telescope, and Keck and Subaru radial velocity measurements, Johnson et al. (2011) report a planetary mass of 0.71 ± 0.03 M_J and radius of 1.34 ± 0.07 R_J, which agree with our values.

The values for stellar density, effective temperature, and metallicity from the MCMC analyses (MS constraint off, eccentricity fixed to zero) were used in an interpolation of the Padova stellar evolution tracks (Girardi et al. 2002; Marigo et al. 2008), shown in Figures 8–10. This resulted in an age for WASP-35 of 5.01 ± 1.16 Gyr and a mass of 1.03 ± 0.04 M_☉.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

WASP-48 was found to have an age of 7.9^+2.0_−1.6 Gyr and mass of 1.08 ± 0.05 M_☉, supporting the results from the MCMC analyses and lack of lithium and Ca H+K that WASP-48 is an old, slightly evolved star. However, the inferred age from the stellar rotation period is still only 0.6^+0.4_−0.2 Gyr when based on a rotational period of 7.2 ± 0.5 days found from the larger stellar radius of 1.75 ± 0.09 R_☉ (with vsin i = 12.2 ± 0.7 km s⁻¹, as before). This apparent contradiction may be due to stellar spin-up by the planet. The close orbits of most known transiting exoplanets produce strong tidal forces between the planet and host star, which lead eventually to orbital circularization, synchronization, and decay. There is evidence for tidal circularization among the known transiting exoplanets, seen in the period–eccentricity relation where planets with very short orbital periods have circular orbits while those with longer periods show a range of eccentricities, but the synchronization of stellar rotation is generally ruled out for most known planet–star systems, with timescales of orders of magnitude larger than the Hubble time (Mazeh 2008; Pont 2009). A few exceptions exist where the companion is massive, e.g., the unique case of τ Boo for which the planetary orbital period coincides closely with the mean rotation period of the stellar differential rotation pattern determined from the Zeeman–Doppler imaging (Fares et al. 2009). WASP-48b orbits with a period of 2.14 days (a = 0.034 AU), and while it is not a massive companion, the evolved nature of the star giving a radius of 1.75 R_☉ results in a relatively short stellar tidal synchronization time of 11⁺³₋₂ Gyr. A full tidal analysis will be discussed in the forthcoming paper of D. Brown (2011, in preparation).

For HAT-P-30/WASP-51, the interpolation yields an age of 4.0^+1.0_−0.8 Gyr and mass of 1.16 ± 0.04 M_☉.

An alternative analysis was also performed using the Yonsei-Yale isochrones (Demarque et al. 2004), which produced an age estimate of 4⁺²₋₁ Gyr and a mass estimate of 1.04 ± 0.04 M_☉ for WASP-35, 6.0^+2.0_−1.0 Gyr and 1.14^+0.06_−0.07 M_☉ for WASP-48, and 3.0^+1.0_−0.5 Gyr and 1.20^+0.03_−0.05 M_☉ for HAT-P-30/WASP-51.

The results of both isochrone analyses are consistent within uncertainties, although the Padova interpolations yield higher ages for each system than the Yonsei-Yale interpolations, and mass values agree with the calibrated values from the MCMC analysis of 1.08 ± 0.03 M_☉, 1.19 ± 0.05 M_☉, and 1.18 ± 0.03 M_☉ for WASP-35, WASP-48, and HAT-P-30/WASP-51, respectively.