EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 553 (May 2013) A&A, 553 (2013) A26 Full HTML
Free Access
Issue
A&A
Volume 553, May 2013
Article Number A26
Number of page(s) 8
Section Planets and planetary systems
DOI https://doi.org/10.1051/0004-6361/201321084
Published online 25 April 2013

© ESO, 2013

1. Introduction

The discovery and the detailed analysis of planets orbiting other stars can yield much information about their physical properties and evolutionary pathways. Since the first detection of a Jupiter-like planet hosted by a main sequence star (Mayor & Queloz 1995), several hundred exoplanets have been discovered via different methods. These comprise radial velocity (RV), astrometry, transits, microlensing, direct imaging and timing (see Schneider et al. 2011). Among the currently known exoplanets, those that transit their host stars command great interest because so many of their physical properties can be determined (Seager & Mallén-Ornelas 2003; Torres et al. 2010; Sozzetti et al. 2007). The properties of particular value include their radii and masses, and hence their surface gravities and mean densities. Observations of planetary transits currently provide the best route to constructing the mass−radius diagram of exoplanets, which brings critical information on the structure, formation and evolution of these worlds (Baraffe et al. 2010; Fortney et al. 2007; Burrows et al. 2007). Moreover, accurate measurements of the times of transit midpoints can allow the detection of additional bodies in known exoplanetary systems (Holman & Murray 2005) or alternatively may serve as a tool to verify the planetary nature and study the dynamical effect in multi-planet systems (Lissauer et al. 2011).

Notably, exoplanetary transits provide an opportunity to study the atmospheres of these objects. During transits part of the star light filters through the planetary atmosphere, and is imprinted with characteristic signatures of atomic (such as Na and K) and molecular (e.g. H2O, CO and CH4) absorption (Seager & Sasselov 2000; Hubbard et al. 2001).

The amount of irradiation incident upon giant extrasolar planets is an important factor in determining the properties of their atmospheres. One theory of giant exoplanets at high incident stellar fluxes predicts the existence of the pM and pL classes of hot Jupiters, depending on the presence of strong absorbers, such as gaseous TiO and VO in their atmospheres (Fortney et al. 2008, 2010). In particular, the increased opacity due to TiO and VO in the pM class should result in about a 3% variation of the observed radius of a planet with a surface gravity of 15 m s-2, over the wavelength interval 350–700 nm. For the purposes of transmission spectroscopy it is practical to convert these planet variations to the change of a directly observed quantity such as the transit depth δ = (Rp/R∗)2. Assuming a radius variation with an amplitude n = 0.03 (3%), the transit depth variation is defined as (1)A transiting Jupiter orbiting a Sun-like star causes δ = 0.01 which would translate into a variation of Δδ = 6 × 10-4 (or ). Such precisions may be achieved by modelling light curves obtained using ground-based 2–4 m telescopes. The best results can be obtained by defocusing the telescope, which averages out sensitivity variations between the pixels on the detector and improves the observing efficiency by enabling a larger fraction of open-shutter time. However, the bandwidths of the filters used (which give the spectral resolution) is the other critical factor affecting the chances of detecting the signal in Eq. (1), similar to the detection of spectral features in spectrophotometry.

Several successful searches for the theoretically predicted strong absorbers in the optical and near-infrared (NIR) have been conducted first from space using the Hubble and Spitzer space telescopes, resulting in the detection of Na, CO, H2O, CH4 and Rayleigh scattering in the blue (Charbonneau et al. 2002; Pont et al. 2008; Sing et al. 2011b). Hampered by the atmosphere of the Earth (e.g. telluric contamination) and instrumental systematics, the ground-based detection of spectral features in exoplanet atmospheres took longer to yield results (Redfield et al. 2008; Snellen et al. 2008; Sing et al. 2011a, 2012).

thumbnail Fig. 1

Simultaneous optical to near-infrared (Sloang′, r′, i′, ′z and Johnson J,H,K) raw transit light curves of WASP-23 (dots), obtained with GROND on UT February 25, 2012 an the corresponding correction functions (red lines).

Although transit photometry or spectrophotometry is often of high quality, the method pushes the limits of current instrumentation and has yielded controversial results (Swain et al. 2008; Beaulieu et al. 2011). The success of the method may also be hampered by the activity of the host star, in particular spot crossing events which can significantly increase the complexity of light curve analysis and lower the precision of the planetary radius measurement (Knutson et al. 2011).

The transiting hot Jupiter WASP-23 b orbits a moderately bright (V = 12.68) K1 V (Teff = 5150 ± 100 K) star and was found by the SuperWASP survey (Triaud et al. 2011; Pollacco et al. 2006). The planet (radius 0.96 ± 0.05  RJ, mass 0.88 ± 0.10   MJ) has a 2.94 d circular orbit (e < 0.062 at the 3σ level).

In this paper we present ground-based simultaneous optical and NIR photometry of a transit of WASP-23 b. We refine its measured properties and orbital ephemeris. We then construct the first transmission spectrum of WASP-23 b and investigate the variation in radius of the planet with wavelength. The layout of the paper is as follows. In Sect. 2 we present the observations. Sect. 3 discusses our data reduction methods. In Sect. 4 we report the analysis of the data, measure the physical properties of the system, construct the planet’s transmission spectrum, and compare this theoretical predictions. We conclude in Sect. 5.

2. Observations

We monitored the flux of WASP-23 (also known as GSC 07635-01376, α = 06h44m30.65s, δ = −42°45′41.0′′; J2000.0; V = 12.68 mag) during one transit on February 25, 2012 (Figs. 1 and 2). The data were collected using the GROND instrument, attached to the MPG/ESO 2.2 m telescope at ESO La Silla Observatory (Chile)1. GROND (Gamma Ray Burst Optical and Near-Infrared Detector) is a simultaneous multi-channel imager that has been specifically designed for gamma-ray burst (GRB) afterglow observations (Greiner et al. 2008). Its capability to simultaneously monitor sources in seven optical to NIR passbands make it an excellent instrument for obtaining high quality, multi-band transit light curves. It is capable of yielding light curves with a root-mean-square (rms) scatter better than 1 mmag, depending on the atmospheric conditions (Lendl et al. 2010; Nikolov et al. 2012; Mancini et al. 2013a).

thumbnail Fig. 2

Left panel: detrended transit light curves (dots) of and the best-fit models (lines) for the WASP-23 GROND data. The optical photometry clearly exhibits a change in the shape of the transit due to the different limb-darkening in each band. Right panel: residuals of the light curve compared to the fitted models.

To reach GROND’s detectors the light from the telescope is split using dichroics into seven beams, resulting in response curves close to the Sloan g′ (λ = 459 nm), r′ (λ = 622 nm), i′ (λ = 764 nm), z′ (λ = 899 nm) in the optical and J (λ = 1256 nm), H (λ = 1647 nm) and K (λ = 2151 nm) bands in the NIR, respectively. The optical beams are incident onto four 2048   × 2048 back-illuminated E2V CCDs with 5.4′   ×   5.4′ fields of view (FOVs) at a plate scale of pixel-1. The NIR channels are imaged onto three 1024 × 1024 Rockwell HAWAII-1 arrays with 10 × 10 arcmin FOVs and plate scales 06 pixel-1. The off-axis guide camera was used to autoguide the telescope.

Light curves were acquired in GROND’s non-dither mode to decrease systematics associated with inter-pixel sensitivity variations. We obtained data for WASP-23 and eleven and seven reference stars in the optical and NIR images, respectively. We applied a moderate defocus to the 2.2 m telescope, such that the point spread functions (PSFs) had diameters ~76 pixels (12″) in the optical and ~13 pixels () in NIR. Table 1 gives details of the instrumental setup and illumination levels of the seven beams. As the guide camera was also defocussed, we made sure in advance that at least one bright star was within its FOV.

Table 1

Instrumental and observational settings used to obtain data for WASP-23 (median values), during the observation on UT February 25, 2012.

The main advantages of the telescope defocus technique have been discussed by several researchers, e.g. Winn et al. (2007), Gillon et al. (2007), Southworth et al. (2010, and references therein), Croll et al. (2010). Although the method results in non-uniformly defocused PSFs of the stars across large FOVs (≥1° × 1°), the technique performs well on small FOVs (such as GROND). It allows superb PSF sampling, significantly reducing the random and systematic errors resulting from detector imperfections and response variations. The technique also avoids detector saturation due to fluctuations in seeing, because such fluctuations are small compared to the level of defocusing. We note that telescope defocusing is not well suited to photometry in crowded fields or when the target star has nearby companions. We used the 2MASS catalogue to rule out these possibilities for our observations of WASP-23.

We observed WASP-23 on February 25, 2012 for 4.4 h from UT 00h32m to 04h56m. An exposure time of 90 s was used in the optical and resulted in 91 integrations. The slow read-out mode was selected (~48 s read-out-time) as it delivers high-quality images2. In the NIR averaged stacks of seven 10 s integrations were obtained to improve the signal-to-noise ratio (S/N), resulting in a total of 181 images. At the beginning of the observation the airmass of the target began at 1.04, peaked at 1.03, and increased to 1.47 by the end of the observing sequence. The observing conditions were nearly photometric, with seeing variation from ~16 to ~18 pixels as measured from a fit to the wings of the defocused PSFs.

Theoretical models of exoplanet transmission spectra imply that the NIR region is richer than the optical in terms of potential spectroscopic signatures. We therefore aimed to obtain high-precision light curves for that region as well. The control of GROND allows only a single integration time (limited from 0 to 10 s) for the three JHK bands simultaneously, so we optimised the exposure time such that the highest S/N was obtained for the band when the star had the most flux (J). For the degree of telescope defocus, we aimed to optimise the S/N in the optical, as we expected better performance in that region based on previous experience (Nikolov et al. 2012; Mancini et al. 2013a).

3. Data reduction and analysis

The data were reduced and analysed on a channel-by-channel basis, using a customised pipeline. GROND’s optical and NIR data are provided in separate fits files. The images from the optical channels are packed into four-extension data cubes and the JHK read-outs are delivered as a single 3072 × 1024 image. After processing to collapse the optical data cubes and to split each NIR image, the data were sorted into science and calibration images.

For the optical data a median-combined master bias frame was computed using 20 zero-second images. We estimated the average dark current of GROND’s optical detectors to be ~0.01  e−   s-1   pixel-1, which is negligible for our data. We therefore neglected the dark current correction. A median-combined master flat-field was calculated using six bias-corrected and dithered twilight sky flat-fields that survived a selection criterion for linearity in each of the passbands. Finally, each optical science frame was de-biased and flat-field corrected to produce calibrated time series of images.

In the NIR, a series of ten dark frames with integration time of 10 s were employed to compute a median-combined master dark frame. A series of 24 dithered twilight sky flat frames were used to obtain a median-combined master sky flat. The master dark frame was subtracted from all the science and sky flat frames. Each dark-subtracted science image was then corrected with the master sky flat frame. An electronic odd-even readout pattern along the y-axis was removed before flat-fielding. This was done by smoothing each image and comparing it with the unsmoothed one after the master dark had been subtracted. The amplitudes of the read-out pattern were then determined by comparing the median level of each column to the overall median level. Each column was then shifted to the overall median level. The master sky flats were divided out from the science images after removal of the read-out pattern.

Previous investigators have commonly obtained dithered images prior to the main observing sequence to further correct the science images against sky background variations. However, similar corrections are inadequate for sky gradients that vary on timescales less than the duration of the observations. In addition, relative photometry, if performed at gradient-stable sky backgrounds is in practice insensitive to sky variations that concern simply the sky level. Finally, previous analyses of WASP-44 b (Mancini et al. 2013a) with and without sky corrections resulted in indistinguishable light curves with similar scatter. We therefore did not obtain dithered sky images.

Table 2

r′-band light curve extract.

Table 3

Theoretical limb-darkening coefficients, light curve residual scatter (in parts per million, ppm) and the β-factor used to estimate the photometric uncertainties of the GROND data.

We performed circular aperture photometry on each of the calibrated images. To accurately determine the central positions of the heavily defocused PSFs, we proceeded as follows. First, we extracted sub-images, roughly centred on each star. The brightest stars were rings with typical diameters ~50 and ~15 pixels in the optical and NIR, respectively. We then convolved each image with a Gaussian kernel to produce a smoothed PSF with a well-defined peak, located at the centre of the original PSF. Finally, we fitted a two-dimensional Gaussian to measure the PSF centre and full width at half maximum, which was latter used to monitor the seeing variations during the observation. Instrumental fluxes and uncertainties (computed taking into account a random photon noise) were acquired using a daophot-type photometry algorithm as implemented in the idl3aper4 routine.

To compute the relative flux light curve of WASP-23 in the four optical channels we used 2MASS J06444546-4242083 as a comparison star. Because of the relatively small FOV of GROND we performed an exhaustive investigation of the possible reference stars prior to the observations. This object satisfied the requirements, being isolated enough to avoid contamination with bright neighbours when defocused, and being the only stellar source in the optical FOV of a similar brightness and colour to WASP-23, thus reducing systematic effects due to differential colour extinction. For instance, the instrumental magnitude differences (WASP-23 minus comparison star), measured on UT Februray 25 2012 were Δg′ = 0.510 mag, Δr′ = 0.082 mag, Δi′ = 0.319 mag, Δz′ = 0.785 mag. For the same reasons, in the NIR channels we used 2MASS J06445925-4245061 as the comparison star. The instrumental magnitude differences were ΔJ = −0.784 mag, ΔH = −0.696 mag and ΔK = −0.396 mag.

To select optimum sizes for the inner aperture and sky annulus we investigated the root-mean-square (rms) of the out-of transit baseline flux as a function of various combinations of aperture and annulus sizes. Table 2 exhibits an r′-band light curve extract along with the airmass and measurements of the (x,y) drift of the PSF on the detector employed to correct the light curve during the analysis.

We used the UT time stamps available in the fits headers of the data and computed the central times of each integration using the exposure times. All times were then converted to Barycentric Julian Date (BJD), using the converter of Eastman et al. (2010).

4. Light curve analysis

We performed the light curve analysis in two steps. First, we simultaneously fitted the seven optical to NIR GROND light curves to derive accurate estimates of the transit central time Tc and orbital period P, and evaluated the orbital parameters and planetary radius (i, a/R∗, Rp/R∗).

In the second step we complemented the GROND photometry with the radial velocity (RV) measurements of WASP-23, reported by Triaud et al. (2011) and performed a joint photometry plus RV MCMC analysis using exofast (Eastman et al. 2013). As our ultimate goal is to construct a ground-based transmission spectrum of WASP-23, based on uniform transit parameters from all bands we excluded the RVs obtained during transit, which are affected by the Rossiter-McLaughlin effect.

As exofast allows an analysis of one light curve at a time we use the results from the first step (the simultaneous fit to the photometry) as priors for exofast. The latter was used to derive reliable estimates of the physical properties and their uncertainties for WASP-23 b and its parent star. exofast utilizes empirical polynomial relations (Torres et al. 2010) between the masses and radii of stars, and their surface gravities (log g), effective temperatures (Teff) and metallicities [Fe/H], based on a large sample of well-studied non-interacting binaries. Including this empirical relation as a constraint allows a derivation of the full orbital and physical properties from transit light curves and RV data (Eastman et al. 2013).

In the first step, we modelled the transit light curves simultaneously in flux units). A two-component correction function was constructed, containing a transit model multiplied by a baseline function that accounts for the most significant known systematic effects affecting ground-based transit observations: airmass and detector intra-pixel sensitivity variations (Fig. 1).

The first component is based on the analytical transit models of Mandel & Agol (2002), which in addition to the transit central time (Tc) and orbital period (P) is a function of the orbital inclination (i), planet semi-major axis and radius normalized to the stellar radius (a/R∗) and (Rp/R∗), respectively. To account for the limb darkening of the star we adopted the quadratic limb darkening law: (2)where I is the intensity and μ is the cosine of the angle between the line of sight and the normal to the stellar surface. We fixed the two limb darkening coefficients (u1 and u2) to their theoretical values (see Table 3 for details), which we obtained by interpolating within the calculated and tabulated ATLAS models (Claret & Bloemen 2011; Eastman et al. 2013) to the stellar parameters Teff = 5150 ± 100 K, log   g = 4.4 ± 0.2 (cgs), [Fe/H]  = −0.05 ± 0.13 and vt = 0.8 ± 0.3 km s-1, which we adopted from Triaud et al. (2011).

We constructed a baseline function to account for systematic effects that concern the (in- and out-of-transit) photometry. It contains a linear airmass (z) term as well as a quadratic polynomial of the PSF location (x,y) on the detector to take into account for any inter-pixel sensitivity variations: (3)To derive the best-fit parameters we employed the Levenberg-Marquardt least-squares algorithm, minimising the χ2 function over the data (Fig. 2) (Markwardt 2009). Conveniently, exofast also takes into account the systematics, incorporating a baseline function and trends, as described in Eastman et al. (2013). We therefore included function (3) in exofast to take into account the systematics and to enable a uniform analysis of the in- and out- of-transit photometry5.

An accurate derivation of system parameters from transit light curve fits requires a realistic estimation of the related photometric errors. The photometric uncertainties produced by idl/aper are usually underestimates, so we ignored them in our analysis. These are formal statistical uncertainties and do not take into account time-correlated noise. We used the out-of-transit data of each light curve before detrending and applied the “time-averaging” procedure proposed by Pont et al. (2006) and used by various authors including Gillon et al. (2006), Winn et al. (2007, 2008), Gibson et al. (2008), Nikolov et al. (2012), Southworth et al. (2012a,b) and Mancini et al. (2013b) to inflate the photometric errors to more realistic values. In summary, for each band we computed the β-ratio between the standard deviation of the out-of-transit unbinned and binned data with bin sizes similar to the transit ingress/egress duration. The photometric uncertainties were therefore set to the out-of-transit unbinned standard deviation multiplied by the corresponding β-factor (see Table 3).

4.1. System parameters

To derive system parameters for WASP-23, including the priors for exofast, we relied on the light curves from the seven GROND passbands. We assumed that i and a should not depend on the observed band and constrained them to a single universal value. In contrast, we treated the planetary radius as a free parameter for each passband. We also set the linear and the quadratic limb darkening coefficients to their theoretical values as well as the initial guess for the orbital period to the value reported in Triaud et al. (2011). We found i = 88.02° ± 0.19°, a/R∗ = 9.94 ± 0.11 and ratios of the radii for each passband: Rp/R∗(g′) = 0.13410 ± 0.00066, Rp/R∗(r′) = 0.13348 ± 0.00049, Rp/R∗(i′) = 0.13455 ± 0.00047, Rp/R∗(z′) = 0.13439 ± 0.00048, Rp/R∗(J) = 0.1340 ± 0.0016, Rp/R∗(H) = 0.1352 ± 0.0016, Rp/R∗(K) = 0.1325 ± 0.0020. These values were input in exofast as priors along with the star’s Teff, log g and [Fe/H], fitting the transit photometry and RVs simultaneously in each band. The results for the main physical properties are summarised in Tables 4 and 5. Our results are slightly different from these reported in Triaud et al. (2011), though with higher precision, favouring a more inflated planet. In particular we improved the precision of i by factors of ~4 and ~6.5 for the lower and upper limits, respectively. We find the radius of WASP-23b to be , which is larger than the radius presented in the discovery paper.

Table 4

Derived system parameters for WASP-23 b from light curve fits, compared to the corresponding results reported in the discovery paper.

4.2. Ephemeris

One of our main goals was to secure the transit ephemeris (Tc and P) of WASP-23 b. We fixed all physical parameters to their values derived in the previous section and fitted GROND’s light curves simultaneously to derive a transit time. We added this to all reported central transit times in the literature (Table 6), then performed a linear fit as a function of the observed epoch (E): (4)We derived P = 2.9444300 ± 0.0000011 d and Tc = 2   454   813.68159 ± 0.00042 BJD. This period is in poor agreement (~3σ level) with the value reported in Triaud et al. (2011) ( d), but is more reliable due to the larger time span of transit times we used. We investigated the observed minus calculated (O–C) residuals of the timings versus our ephemeris. We found no systematic deviation and conclude that there is no compelling evidence for transit timing variations (Fig. 3).

Table 5

Stellar and planetary parameters for WASP-23 derived from light curve and RV fits to the GROND photometry and RV measurements from Triaud et al. (2011).

4.3. Radius variation

A set of simultaneous multi-band transit light curves provides an opportunity to investigate the behaviour of the measured planetary radius with wavelength. Atmospheric absorption features in the optical regime due to sodium (Na) and potassium (K) have been theoretically predicted (Seager & Sasselov 2000) and observed in the transmission spectra of hot Jupiters, e.g. XO-2b (Sing et al. 2011b, 2012) from the ground. Contrary to this example of an agreement between theory and observation, an analysis of HST ultraviolet-optical spectrophotometric data of HD 189733 b by Sing et al. (2011a) revealed a featureless spectrum, which might be an indication for the presence of high altitude clouds. These observational results indicate the diversity of exoplanet atmospheres, the understanding of which requires more observational results.

To investigate the radius variation of WASP-23 b in GROND’s passbands we fixed P, i and a/R∗ to their final results from the analysis in Sect. 4.1, and the linear and quadratic limb darkening coefficients to their predicted values. We fitted for the planetary radius (specifically the ratio of the radii Rp/R∗) and the instrumental systematics. This approach improves the precision of the measured planetary radius, by avoiding sources of uncertainty common to all measurements of Rp/R∗. In addition, the fit takes into account the information for the system parameters and reflects differences in the transit depth that are determined only by the size of the planet in each passband. To estimate realistic values for the normalised planet radii and their error bars we carried out the fitting using an MCMC approach. The results for the planet radii, obtained at this step are displayed in Figs. 4 and 5. The vertical error bars comprise the 68.3% confidence levels from the MCMC distributions of Rp/R∗. As expected, the radius measurements in the optical regime show much smaller error bars compared to the NIR passbands. This is due to the lower scatter of the data in the optical. We also show horizontal bandwidths (plotted with error-like bars) that have been computed as the full widths at half maximum, using GROND’s filter curves, which are indicated at the lower part of the plot.

To compare our observational results with theorical predictions we employed the system parameters from Table 5 and computed a 1D model atmosphere of WASP-23 b, using the code described in Fortney et al. (2005, 2008). The fully non-gray model uses the chemical equilibrium abundances of Lodders & Fegley (2002) and the opacity database described in Freedman et al. (2008). The atmospheric pressure-temperature profile simulates planet-wide average conditions. We computed the transmission spectrum of the model using the methods described in Fortney et al. (2010). For each of the GROND passbands we computed the expected theoretical value given the synthetic transmission spectrum, integrating over the corresponding filter curve. Finally, we used the derived theoretical values for the planet radius and fitted them as a model to the measured values of Rp/R∗ from the seven GROND light curves (Figs. 4 and 5). We find a good agreement between the observationally derived measurements and synthetic values of the planetary radii, consistent with a featureless transmission spectrum at the ~1σ level. Two of our radius estimates (r′-band in the optical and H-band in the NIR) were found to be off by slightly more than 1σ from their corresponding theoretical values. This is most likely due to systematics of unknown origin.

5. Summary and discussion

We have presented observational results based on data collected with the GROND instrument on the MPG/ESO 2.2 m telescope at ESO La Silla, during one transit of the hot Jupiter WASP-23 b, the first photometric follow-up of this object after its discovery. We employed the telescope defocusing technique and autoguiding to optimise the quality of the light curve and minimise the instrumental systematics in the complete optical to NIR capability of the GROND instrument for the first time. Light curves constructed from this data set showed a scatter versus fitted models as low as 330 ppm (Table 3). For comparison, a recent observation performed by Tregloan-Reed & Southworth (2013) with the 3.58 m New Technology Telescope (NTT), in defocussed mode obtained light curve scatters of 258 and 211 ppm, during two transit events of WASP-50 b. Their lower scatter is expected due to the larger aperture of the telescope.

Table 6

Transit mid-times and O–C residuals computed using transit light curves from the literature and the observation reported in this work.

thumbnail Fig. 3

Observed minus computed (O–C) transit times for WASP-23 b, using observations reported in the literature (black dots) and our measurement (orange dot), compared to a linear ephemeris (dashed line).

thumbnail Fig. 4

Radius variation (measured with Rp/R∗ and indicated with yellow disks with error bars) of WASP-23 b in the seven optical to NIR passbands, compared to a synthetic transmission spectrum based on a planet-wide pressure-temperature profile (grey continuous line). The open red diamonds indicate the expected theoretical values, integrated over GROND’s filter curves. The latter are displayed as dotted lines at the origin of the graph.

thumbnail Fig. 5

Same as Fig. 4, but displaying a closer view of the radius variation in the optical bands.

We used the full potential of GROND’s optical to NIR dataset to refine the ephemeris and system parameters of the WASP-23b system and compared our results with these presented by Triaud et al. (2011). We complemented our photometric data with existing RV measurements of the parent star to derive the full physical properties of the system. We measured the central transit time with a precision ~8 s, complemented this measurement with published transit epochs, and refined the orbital period of WASP-23 b. We analysed the O–C diagram of the transit times and found no significant evidence for transit timing variations. Using all of the seven GROND bands we determined a larger and more precise value for the radius of WASP-23 b, which is in agreement at the ~2σ level with that from Triaud et al. (2011). The latter however, was derived based on a joint fit to five lower-quality light curves covering the z, R and I + z passbands. Despite our larger value for the planetary radius, WASP-23 b still sits below the region of the so-called heavily “bloated” hot Jupiters, in accord to the theoretical models for irradiated giant planets of Baraffe et al. (2008). Moreover, Fortney et al. (2007) presented theoretical mass-radius relationships for irradiated giant exoplanets, parametrized by mass, age, and an effective orbital distance, defined as the distance from the Sun, where a hypothetical planet would receive the same flux as the actual planet. Given the luminosity of WASP-23 and the semi-major axis of its planet we evaluated that distance to be a = 0.585 AU. We then interpolated the models of Fortney et al. (2007) to match the distance and the mass of WASP-23b for a solar composition at 4.5 Gyr and evaluated Rp = 1.042  RJup, which is within the error of our result.

Finally, we used the potential of the GROND light curves to investigate the (possible) variation of the planetary radius with wavelength and constructed a ground-based simultaneous optical to NIR broad-band transmission spectrum of WASP-23 b. According to our results for the system parameters (Table 5) the equilibrium temperature of WASP-23 b is  K, which places the planet among the pL-class (cooler) of giant close-in exoplanet atmospheres in the scheme proposed by Fortney et al. (2008). Taking into account these results, we generated a synthetic atmospheric transmission model of an irradiated hot Jupiter, assuming a planet-wide pressure-temperature profile and compared theory with observations. The optical part of the synthetic transmission spectrum at this equilibrium temperature and surface gravity (20.73 m s-2) suggests a radius variation due to the presence of gaseous sodium and potassium in the atmosphere of WASP-23 b, while the NIR parts are dominated by water vapour features. The detection of such radius variation should be a realistic task, if observations could be taken at the precision levels attained in this work and with a higher spectral resolution. However, even the narrowest passband of GROND (i′-band) spans ~120 nm, which effectively reduces the sensitivity of our measurements to radius variations. This effect is illustrated in Figs. 4 and 5, where using the filter curves of GROND we computed the theoretically expected values for the radius of WASP-23 b, given the synthetic spectrum. We fitted our observational result for the radius to the theoretical estimates. Although we found the r′-band, which covers the expected sodium absorption line, to deviate from the predicted theoretical value, overall the theoretical model is in very good agreement with observational transmission spectrum at the ~1σ level.

1

Detailed information on the 2.2 m telescope, the GROND instrument and the relevant technical settings can be found at http://www.eso.org/sci/facilities/lasilla/telescopes/2p2/

2

The GROND detector can also be read-out in “fast” mode (~11 s read-out-time), but it has been found during previous transit observations that images obtained in that mode suffer from systematic effects that lower the quality of the resulting light curves.

3

The acronym IDL stands for Interactive Data Language. For further details see http://www.exelisvis.com/ ProductsServices/IDL.aspx

4

Part of aper Astronomy User’s Library. For detailed information see http://idlastro.gsfc.nasa.gov/

5

We choose to employ mpfit in the first step as it allows one to construct a joint model incorporating the entire dataset. In contrast exofast permits an analysis on a per-channel basis, i.e. one band at a time. We therefore, determined universal values for a/R∗, i, Tc, P and Rp/R∗ using the complete dataset and used these results as priors in exofast, where the fitting is done for each band.

Acknowledgments

The authors acknowledge Timo Anguita and Régis Lachaume for their technical support at the time of the observations, David Sing and Frédéric Pont for their constructive discussions as well as the anonymous referee for the detailed and constructive suggestions that improved the manuscript. N.N. acknowledges support from an STFC consolidated grant. J.S. acknowledges STFC for the award of an Advanced Fellowship.

References

  1. Baraffe, I., Chabrier, G., & Barman, T. 2008, A&A, 482, 315 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  2. Baraffe, I., Chabrier, G., & Barman, T. 2010, Rep. Prog. Phys., 73, 016901 [NASA ADS] [CrossRef] [Google Scholar]
  3. Beaulieu, J.-P., Tinetti, G., Kipping, D. M., et al. 2011, ApJ, 731, 16 [NASA ADS] [CrossRef] [Google Scholar]
  4. Brogi, M., Snellen, I. A. G., de Kok, R. J., et al. 2012, Nature, 486, 502 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  5. Burrows, A., Hubeny, I., Budaj, J., & Hubbard, W. B. 2007, ApJ, 661, 502 [NASA ADS] [CrossRef] [Google Scholar]
  6. Charbonneau, D., Brown, T. M., Noyes, R. W., & Gilliland, R. L. 2002, ApJ, 568, 377 [NASA ADS] [CrossRef] [Google Scholar]
  7. Claret, A., & Bloemen, S. 2011, A&A, 529, A75 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  8. Croll, B., Albert, L., Lafreniere, D., et al. 2010, ApJ, 717, 1084 [NASA ADS] [CrossRef] [Google Scholar]
  9. Eastman, J., Silverd, R., & Gaudi, B. S. 2010, PASP, 122, 935 [NASA ADS] [CrossRef] [Google Scholar]
  10. Eastman, J., Gaudi, B. S., & Agol, E. 2013, PASP, 125, 83 [NASA ADS] [CrossRef] [Google Scholar]
  11. Fortney, J. J., Marley, M. S., Lodders, K., Saumon, D., & Freedman, R. 2005, ApJ, 627, L69 [NASA ADS] [CrossRef] [Google Scholar]
  12. Fortney, J. J., Marley, M. S., & Barnes, J. W. 2007, ApJ, 659, 1661 [NASA ADS] [CrossRef] [Google Scholar]
  13. Fortney, J. J., Lodders, K., Marley, M. S., & Freedman, R. S. 2008, ApJ, 678, 1419 [NASA ADS] [CrossRef] [Google Scholar]
  14. Fortney, J. J., Shabram, M., Showman, A. P., et al. 2010, ApJ, 709, 1396 [NASA ADS] [CrossRef] [Google Scholar]
  15. Freedman, R. S., Marley, M. S., & Lodders, K. 2008, ApJS, 174, 504 [NASA ADS] [CrossRef] [Google Scholar]
  16. Gibson, N. P., Pollacco, D., Simpson, E. K., et al. 2008, A&A, 492, 603 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  17. Gillon, M., Pont, F., Moutou, C., et al. 2006, A&A, 459, 249 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. Gillon, M., Pont, F., Demory, B.-O., et al. 2007, A&A, 472, L13 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  19. Greiner, J., Bornemann, W., Clemens, C., et al. 2008, PASP, 120, 405 [NASA ADS] [CrossRef] [Google Scholar]
  20. Hansen, B. M. S., & Barman, T. 2007, ApJ, 671, 861 [NASA ADS] [CrossRef] [Google Scholar]
  21. Holman, M. J., & Murray, N. W. 2005, Science, 307, 1288 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  22. Hubbard, W. B., Fortney, J. J., Lunine, J. I., et al. 2001, ApJ, 560, 413 [NASA ADS] [CrossRef] [Google Scholar]
  23. Knutson, H. A., Madhusudhan, N., Cowan, N. B., et al. 2011, ApJ, 735, 27 [NASA ADS] [CrossRef] [Google Scholar]
  24. Lendl, M., Afonso, C., Koppenhoefer, J., et al. 2010, A&A, 522, A29 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  25. Lissauer, J. J., Fabrycky, D. C., Ford, E. B., et al. 2011, Nature, 470, 53 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  26. Lodders, K., & Fegley, B. 2002, Icarus, 155, 393 [NASA ADS] [CrossRef] [Google Scholar]
  27. Mancini, L., Nikolov, N., Southworth, J., et al. 2013a, MNRAS, in press [arXiv:1301.3005] [Google Scholar]
  28. Mancini, L., Southworth, J., Ciceri, S., et al. 2013b, A&A, 551, A11 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  29. Mandel, K., & Agol, E. 2002, ApJ, 580, L171 [NASA ADS] [CrossRef] [Google Scholar]
  30. Markwardt, C. B. 2009, Astronomical Data Analysis Software and Systems XVIII, Quebec, Canada, eds. D. Bohlender, P. Dowler, & D. Durand (San Francisco: ASP), ASP Conf. Ser., 411, 251 [Google Scholar]
  31. Mayor, M., & Queloz, D. 1995, Nature, 378, 355 [NASA ADS] [CrossRef] [Google Scholar]
  32. Nikolov, N., Henning, T., Koppenhoefer, J., et al. 2012, A&A, 539, A159 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Pollacco, D. L., Skillen, I., Collier Cameron, A., et al. 2006, PASP, 118, 1407 [NASA ADS] [CrossRef] [Google Scholar]
  34. Pont, F., Zucker, S., & Queloz, D. 2006, MNRAS, 373, 231 [NASA ADS] [CrossRef] [Google Scholar]
  35. Pont, F., Knutson, H., Gilliland, R. L., Moutou, C., & Charbonneau, D. 2008, MNRAS, 385, 109 [NASA ADS] [CrossRef] [Google Scholar]
  36. Redfield, S., Endl, M., Cochran, W. D., & Koesterke, L. 2008, ApJ, 673, L87 [NASA ADS] [CrossRef] [MathSciNet] [Google Scholar]
  37. Schneider, J., Dedieu, C., Le Sidaner, P., Savalle, R., & Zolotukhin, I. 2011, A&A, 532, A79 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  38. Seager, S. 2011, Exoplanets (Tucson, AZ: University of Arizona Press) [Google Scholar]
  39. Seager, S., & Mallén-Ornelas, G. 2003, ApJ, 585, 1038 [NASA ADS] [CrossRef] [Google Scholar]
  40. Seager, S., & Sasselov, D. D. 2000, ApJ, 537, 916 [NASA ADS] [CrossRef] [Google Scholar]
  41. Sing, D. K., Pont, F., Aigrain, S., et al. 2011a, MNRAS, 416, 1443 [NASA ADS] [CrossRef] [Google Scholar]
  42. Sing, D. K., Désert, J.-M., Fortney, J. J., et al. 2011b, A&A, 527, A73 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  43. Sing, D. K., Huitson, C. M., Lopez-Morales, M., et al. 2012, MNRAS, 426, 1663 [NASA ADS] [CrossRef] [Google Scholar]
  44. Snellen, I. A. G., Albrecht, S., de Mooij, E. J. W., & Le Poole, R. S. 2008, A&A, 487, 357 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Southworth, J., Mancini, L., Novati, S. C., et al. 2010, MNRAS, 408, 1680 [NASA ADS] [CrossRef] [Google Scholar]
  46. Southworth, J., Bruni, I., Mancini, L., & Gregorio, J. 2012a, MNRAS, 420, 2580 [NASA ADS] [CrossRef] [Google Scholar]
  47. Southworth, J., Mancini, L., Maxted, P. F. L., et al. 2012b, MNRAS, 422, 3099 [NASA ADS] [CrossRef] [Google Scholar]
  48. Sozzetti, A., Torres, G., Charbonneau, D., et al. 2007, ApJ, 664, 1190 [NASA ADS] [CrossRef] [Google Scholar]
  49. Swain, M. R., Vasisht, G., & Tinetti, G. 2008, Nature, 452, 329 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  50. Torres, G., Andersen, J., & Giménez, A. 2010, A&ARv, 18, 67 [Google Scholar]
  51. Tregloan-Reed, J., & Southworth, J. 2013, MNRAS, in press [arXiv:1212.0686] [Google Scholar]
  52. Triaud, A. H. M. J., Queloz, D., Hellier, C., et al. 2011, A&A, 531, A24 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  53. Winn, J. N., Holman, M. J., Henry, G. W., et al. 2007, AJ, 133, 1828 [NASA ADS] [CrossRef] [Google Scholar]
  54. Winn, J. N., Holman, M. J., Shporer, A., et al. 2008, AJ, 136, 267 [NASA ADS] [CrossRef] [Google Scholar]

All Tables

Table 1

Instrumental and observational settings used to obtain data for WASP-23 (median values), during the observation on UT February 25, 2012.

In the text
Table 2

r′-band light curve extract.

In the text
Table 3

Theoretical limb-darkening coefficients, light curve residual scatter (in parts per million, ppm) and the β-factor used to estimate the photometric uncertainties of the GROND data.

In the text
Table 4

Derived system parameters for WASP-23 b from light curve fits, compared to the corresponding results reported in the discovery paper.

In the text
Table 5

Stellar and planetary parameters for WASP-23 derived from light curve and RV fits to the GROND photometry and RV measurements from Triaud et al. (2011).

In the text
Table 6

Transit mid-times and O–C residuals computed using transit light curves from the literature and the observation reported in this work.

In the text

All Figures

thumbnail Fig. 1

Simultaneous optical to near-infrared (Sloang′, r′, i′, ′z and Johnson J,H,K) raw transit light curves of WASP-23 (dots), obtained with GROND on UT February 25, 2012 an the corresponding correction functions (red lines).
In the text
thumbnail Fig. 2

Left panel: detrended transit light curves (dots) of and the best-fit models (lines) for the WASP-23 GROND data. The optical photometry clearly exhibits a change in the shape of the transit due to the different limb-darkening in each band. Right panel: residuals of the light curve compared to the fitted models.
In the text
thumbnail Fig. 3

Observed minus computed (O–C) transit times for WASP-23 b, using observations reported in the literature (black dots) and our measurement (orange dot), compared to a linear ephemeris (dashed line).
In the text
thumbnail Fig. 4

Radius variation (measured with Rp/R∗ and indicated with yellow disks with error bars) of WASP-23 b in the seven optical to NIR passbands, compared to a synthetic transmission spectrum based on a planet-wide pressure-temperature profile (grey continuous line). The open red diamonds indicate the expected theoretical values, integrated over GROND’s filter curves. The latter are displayed as dotted lines at the origin of the graph.
In the text
thumbnail Fig. 5

Same as Fig. 4, but displaying a closer view of the radius variation in the optical bands.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.