HST hot-Jupiter transmission spectral survey: haze in the atmosphere
of WASP-6b
Nikolov, N., Sing, D. K., Burrows, A. S., Fortney, J. J., Henry, G. W., Pont, F., Ballester, G. E., Aigrain, S.,
Wilson, P. A., Huitson, C. M., Gibson, N. P., Désert, J-M., Lecavelier Des Etangs, A., Showman, A. P., Vidal-
Madjar, A., Wakeford, H. R., & Zahnle, K. (2015). HST hot-Jupiter transmission spectral survey: haze in the
atmosphere of WASP-6b.
Monthly Notices of the Royal Astronomical Society
,
447
(1), 463-478.
https://doi.org/10.1093/mnras/stu2433
Published in:
Monthly Notices of the Royal Astronomical Society
Document Version:
Publisher's PDF, also known as Version of record
Queen's University Belfast - Research Portal:
Link to publication record in Queen's University Belfast Research Portal
Publisher rights
© 2014 The Authors.
This work is made available online in accordance with the publisher’s policies. Please refer to any applicable terms of use of the publisher.
General rights
Copyright for the publications made accessible via the Queen's University Belfast Research Portal is retained by the author(s) and / or other
copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated
with these rights.
Take down policy
The Research Portal is Queen's institutional repository that provides access to Queen's research output. Every effort has been made to
ensure that content in the Research Portal does not infringe any person's rights, or applicable UK laws. If you discover content in the
Research Portal that you believe breaches copyright or violates any law, please contact openaccess@qub.ac.uk.
Download date:10. Aug. 2022
MNRAS 447, 463–478 (2015) doi:10.1093/mnras/stu2433
HST hot-Jupiter transmission spectral survey: haze in the
atmosphere of WASP-6b
N. Nikolov,
1‹
D. K. Sing,
1
A. S. Burrows,
2
J. J. Fortney,
3
G. W. Henry,
4
F. Pont,
1
G. E. Ballester,
5
S. Aigrain,
6
P. A. Wilson,
1,7
C. M. Huitson,
8
N. P. Gibson,
9
J.-M. D
´
esert,
8
A. Lecavelier des Etangs,
7
A. P. Showman,
5
A. Vidal-Madjar,
7
H. R. Wakeford
1
and K. Zahnle
10
1
Astrophysics Group, School of Physics, University of Exeter, Stocker Road, Exeter EX4 4QL, UK
2
Department of Astrophysical Sciences, Peyton Hall, Princeton University, Princeton, NJ 08544, USA
3
Department of Astronomy and Astrophysics, University of California, Santa Cruz, CA 95064, USA
4
Tennessee State University, 3500 John A. Merritt Blvd, PO Box 9501, Nashville, TN 37209, USA
5
Lunar and Planetary Laboratory, University of Arizona, Tucson, AZ 85721, USA
6
Department of Physics, University of Oxford, Denys Wilkinson Building, Keble Road, Oxford OX1 3RH, UK
7
CNRS, Institut dAstrophysique de Paris, UMR 7095, 98bis boulevard Arago, F-75014 Paris, France
8
CASA, Department of Astrophysical and Planetary Sciences, University of Colorado, 389-UCB, Boulder, CO 80309, USA
9
European Southern Observatory, Karl-Schwarzschild-Str. 2, D-85748 Garching bei M
¨
unchen, Germany
10
NASA Ames Research Center, Moffett Field, CA 94035, USA
Accepted 2014 November 17. Received 2014 November 17; in original form 2014 July 25
ABSTRACT
We report Hubble Space Telescope optical to near-infrared transmission spectroscopy of
the hot-Jupiter WASP-6b, measured with the Space Telescope Imaging Spectrograph and
Spitzer’s InfraRed Array Camera. The resulting spectrum covers the range 0.29–4.5 µm.
We find evidence for modest stellar activity of WASP-6 and take it into account in the
transmission spectrum. The overall main characteristic of the spectrum is an increasing radius
as a function of decreasing wavelength corresponding to a change of (R
p
/R
∗
) = 0.0071
from 0.33 to 4.5 µm. The spectrum suggests an effective extinction cross-section with a power
law of index consistent with Rayleigh scattering, with temperatures of 973 ± 144 K at the
planetary terminator. We compare the transmission spectrum with hot-Jupiter atmospheric
models including condensate-free and aerosol-dominated models incorporating Mie theory.
While none of the clear-atmosphere models is found to be in good agreement with the data, we
find that the complete spectrum can be described by models that include significant opacity
from aerosols including Fe-poor Mg
2
SiO
4
, MgSiO
3
, KCl and Na
2
S dust condensates. WASP-
6b is the second planet after HD 189733b which has equilibrium temperatures near ∼1200 K
and shows prominent atmospheric scattering in the optical.
Key words: methods: observational – techniques: spectroscopic – planets and satellites: at-
mospheres – stars: activity – planets and satellites: individual: WASP-6b.
1 INTRODUCTION
Multiwavelength transit observations of hot Jupiters provide a
unique window to the chemistry and structure of the atmospheres
of these distant alien worlds. During planetary transits, a small
fraction of the stellar light is transmitted through the planetary at-
mosphere and signatures of atmospheric constituents are imprinted
on the stellar spectrum (Seager & Sasselov 2000). Transmission
E-mail: nikolov.nkn@gmail.com
spectroscopy, where the planet radius is measured as a function
of wavelength, has revolutionized our understanding of extrasolar
gas-giant atmospheres. Numerous studies from both space- and
ground-based observations have led to the characterization and
detection of atomic and molecular features as well as hazes and
clouds in the atmospheres of several hot Jupiters revealing a huge
diversity (Charbonneau et al. 2002; Grillmair et al. 2008; Pont
et al. 2008, 2013; Redfield et al. 2008; Snellen et al. 2008, 2010;
Sing et al. 2011, 2012, 2013; Brogi et al. 2012; Bean et al. 2013;
Birkby et al. 2013; Deming et al. 2013; Gibson et al. 2013a,b; Huit-
son et al. 2013; Wakeford et al. 2013; Nikolov et al. 2014; Stevenson
et al. 2014).
C
2014 The Authors
Published by Oxford University Press on behalf of the Royal Astronomical Society
Downloaded from https://academic.oup.com/mnras/article-abstract/447/1/463/989512
by Queen's University of Belfast user
on 08 February 2018
464 N. Nikolov et al.
Atmospheric hazes have currently been detected in hot-Jupiter
exoplanets over a wide range of atmospheric temperature regimes
including HD 189733b and WASP-12b representing the cool and
hot extremes for hot Jupiters at equilibrium temperatures around
1200 and 3000 K, respectively. Hazes are considered to originate
from dust condensation (forming aerosols in the atmospheres of
exoplanets) or a result from photochemistry (Marley et al. 2013).
Transmission spectroscopy studies may constrain the most prob-
able condensates responsible for these two planets, with iron-free
silicates (e.g. MgSiO
3
) and corundum (Al
2
O
3
) being two prime can-
didates for the above temperature regimes (Lecavelier Des Etangs
et al. 2008a;Singetal.2013). Cloud-free hot-Jupiter atmospheric
models predict sodium and potassium to be the dominant absorbing
features in optical transmission spectra (Seager & Sasselov 2000;
Fortney et al. 2005).
In this paper, we present new results for WASP-6b from a large
Hubble Space Telescope (HST) hot-Jupiter transmission spectral
survey comprising eight transiting planets. The ultimate aim of the
project is to explore the variety of hot-Jupiter atmospheres, i.e.
clear/hazy/cloudy, delve into the presence/lack of TiO/alkali and
other molecular features, probe the diversities between possible
sub-classes and perform comparative exoplanetology. Initial results
from four exoplanets have been presented so far in Huitson et al.
(2013) for WASP-19b, Wakeford et al. (2013)andNikolovetal.
(2013) for HAT-P-1b and Sing et al. (2013) for WASP-12b and
Sing et al. (2015) for WASP-31b. Already a wide diversity among
exoplanet atmospheres is observed. In this paper, we report new
HST optical transit observations with the Space Telescope Imaging
Spectrograph (STIS) instrument and combine them with Spitzer
InfraRed Array Camera (IRAC) broad-band transit photometry to
calculate a near-UV to near-infrared transmission spectrum, capable
of detecting atmospheric constituents. We describe the observations
in Section 2, present the light-curve analysis in Section 3, discuss
the results in Section 4 and conclude in Section 5.
1.1 The WASP-6b system
Discovered by the Wide Angle Search for Planets, WASP-6 b is
an inflated sub-Jupiter-mass transiting extrasolar planet orbiting a
moderately bright V = 11.9 solar-type, mildly metal-poor star (with
T
eff
= 5375 ± 65 K, log g = 4.61 ± 0.07 and [Fe/H]=−0.20 ± 0.09)
located in the southern part of the constellation Aquarius (Gillon
et al. 2009; Doyle et al. 2013). The planet is moving on an orbit with
a period of P 3.6 d and semimajor axis a 0.042 au. The sky-
projected angle between the stellar spin and the planetary orbital
axis has been determined through observations of the Rossiter–
McLaughlin effect by Gillon et al. (2009) indicating a good align-
ment (β = 11
+14
−18
deg) that consequently favours a planet migration
scenario via the spin–orbit preserving tidal interactions with a pro-
toplanetary disc. Dragomir et al. (2011) refined the WASP-6 system
parameters and orbital ephemeris from a single ground-based transit
observation finding good agreement between their results and the
discovery paper. Although Gillon et al. (2009) claimed evidence for
non-zero orbital eccentricity, an analysis with new radial velocity
data from Husnoo et al. (2012) brought evidence for non-significant
eccentricity. Finally, Doyle et al. (2013) refined the spectroscopic
parameters of WASP-6b’s host star.
A cloud-free hot-Jupiter theoretical transmission spectrum pre-
dicts strong optical absorbers, dominated by Na
I and K I absorption
lines and H
2
Rayleigh scattering for a planetary system with the
physical properties of WASP-6b, including an effective planetary
temperature (assuming zero albedo and f = 1/4 heat redistribution)
Figure 1. Typical STIS G430L and G750L spectra (blue and red continuous
lines, respectively). A fringe-corrected spectrum (offset by 5 × 10
3
counts)
is portrayed above the uncorrected red spectrum for comparison.
of 1194 K and surface gravity g 8ms
−2
(Fortney et al. 2008;
Burrows et al. 2010). The atmosphere of WASP-6b has recently
been probed in the optical regime from the ground by Jord
´
an et al.
(2013). The authors found evidence for an atmospheric haze char-
acterized by a decreasing apparent planetary size with wavelength
and no evidence for the pressure-broadened alkali features.
2 OBSERVATIONS AND CALIBRATIONS
2.1 HST STIS spectroscopy
We acquired low-resolution (R = λ/λ = 500–1040) HST STIS
spectra (Proposal ID GO-12473, P.I., D. Sing) during three transits
of WASP-6 b on
UT 2012 June 10 (visit 9) and 16 (visit 10) with
grating G430L (∼2.7 Å pixel
−1
) and 2012 July 23 (visit 21) with
grating G750L (∼4.9 Å pixel
−1
). When combined, the blue and
red STIS data provide complete wavelength coverage from 2900 to
10 300 Å with a small overlap region between them from 5240 to
5700 Å (Fig. 1). Each visit consisted of five ∼96 min orbits, during
which data collection was truncated with ∼45 min gaps due to Earth
occultations. Incorporating wide 52 × 2 arcsec slit to minimize slit
light losses and an integration time of 278 s, a total of 43 spectra
were obtained during each visit. Data acquisition overheads were
minimized by reading-out a reduced portion of the CCD with a size
of 1024 × 128. This observing strategy has proven to provide high
signal-to-noise ratio (S/N) spectra that are photometrically accurate
near the Poisson limit during a transit event (Brown et al. 2001; Sing
et al. 2011, 2013; Huitson et al. 2012). The three HST visits were
scheduled such that the second and third spacecraft orbits occurred
between the second and third contacts of the planetary transit in
order to provide good sampling of the planetary radius while three
orbits secured the stellar flux level before and after the transit.
The data reduction and analysis process is somewhat uniform
for the complete large HST programme and follows the general
methodology detailed in Huitson et al. (2013), Sing et al. (2013)
and Nikolov et al. (2014). The raw STIS data was reduced (bias-,
dark- and flat-corrected) using the latest version of the CALSTIS
1
pipeline and the relevant up-to-date calibration frames. Correction
of the significant fringing effect was performed for the G750L data
using contemporaneous fringe flat frame obtained at the end of the
1
CALSTIS comprises software tools developed for the calibration of STIS
data (Katsanis & McGrath 1998) inside the
IRAF (Image Reduction and
Analysis Facility; http://iraf.noao.edu/) environment.
MNRAS 447, 463–478 (2015)
Downloaded from https://academic.oup.com/mnras/article-abstract/447/1/463/989512
by Queen's University of Belfast user
on 08 February 2018
HST STIS transit spectroscopy of WASP-6b 465
observing sequence and the procedure detailed in (Goudfrooij &
Christensen 1998, Fig. 1).
Due to the relatively long STIS integration time (i.e. 287 s), the
data contain large number of cosmic ray events, which were iden-
tified and removed following the procedures described in Nikolov
et al. (2013). It was found that the total number of pixels affected
by cosmic ray events comprise ∼4 per cent of the total number of
pixels of each 2D spectrum. In addition, we also corrected all pixels
identified by CALSTIS as ‘bad’ with the same procedure, which
together with the cosmic ray identified pixels resulted in a total
of ∼11 per cent interpolated pixels.
Spectral extractions were performed in
IRAF employing the APALL
procedure using the calibrated .flt science files after fringe and
cosmic ray correction. We performed spectral extraction with aper-
ture sizes in the range 6.0–18.0 pixels with a step of 0.2. The best
aperture for each grating was selected based on the resulting low-
est light-curve residual scatter after fitting the white light curves
(see Section 3 for details). We found that aperture sizes 12.0, 12.2
and 8.6 pixels satisfy this criterium for visits 9, 10 and 21, respec-
tively. We then placed the extracted spectra to a common Doppler
corrected rest frame through cross-correlation to prevent sub-pixel
wavelength shifts in the dispersion direction. A wavelength solution
was obtained from the x1d files from CALSTIS. The STIS spectra
were then used to extract both white-light spectrophotometric time
series and custom wavelength bands after integrating the appropri-
ate flux from each bandpass.
The raw STIS light curves exhibit instrumental systematics s imi-
lar to those described by Gilliland, Goudfrooij & Kimble (1999)and
Brown et al. (2001). In summary, the major source of the systemat-
ics is related with the orbital motion of the telescope. In particular,
the HST focus is known to experience quite noticeable variations
on the spacecraft orbital time-scale, which are attributed to ther-
mal contraction/expansion (often referred to as the ‘breathing ef-
fect’) as the spacecraft warms up during its orbital day and cools
down during orbital night (Hasan & Bely 1993, 1994; Suchkov &
Hershey 1998). We take into account the systematics associated
with the telescope temperature variations in the transit light-curve
fits by fitting a baseline function depending on various parameters
(Section 3.0.1).
2.2 Spitzer IRAC data
Photometric data were collected for WASP-6 during two transits
of its planet on
UT 2013 January 14 and 21 with the Spitzer space
telescope (Werner et al. 2004) employing, respectively, the 4.5 and
3.6 µm channels of the IRAC (Fazio et al. 2004)aspartofpro-
gramme 90092 (P.I. J.-M. Desert). Both observations started shortly
before ingress with effective integration times of 1.92 s per image
and were terminated ∼2.2 h after egress resulting in 8 320 images
(see Fig. 2), which were calibrated by the Spitzer pipeline (version
S19.1.0) and are available in the form of Basic Calibrated Data
(.bcd)files.
After organizing the data, we converted the images from flux in
mega-Jansky per steradian (MJy sr
−1
) to photon counts (i.e. elec-
trons) using the information provided in the
FITS headers. In partic-
ular, we multiplied the images by the gain and individual exposure
time (
FITS header key words SAMPTIME and GAIN) and divided
by the flux conversion factor (FLUXCONV). Timing of each im-
age was computed using the UTC-based Barycentric Julian Date
(BJD
UTC
) from the FITS header key word BMJD_OBS, transform-
ing these time stamps into BJD based on the BJD terrestrial time
(TT) using the following conversion BJD
TT
∼ BJD
UTC
+ 66.184 s
Figure 2. Warm Spitzer IRAC 3.6 and 4.5µm photometry (left and right, respectively). Top panels: raw flux and the best-fitting transit and systematics model;
middle panels: detrended light curves and the best-fitting transit models and binned by 8 min; lower panels: light-curve residuals.
MNRAS 447, 463–478 (2015)
Downloaded from https://academic.oup.com/mnras/article-abstract/447/1/463/989512
by Queen's University of Belfast user
on 08 February 2018
466 N. Nikolov et al.
(Eastman, Siverd & Gaudi 2010). This conversion is preferable
as leap seconds are occasionally added to the BJD
UTC
standard
(Knutson et al. 2012; Todorov et al. 2013).
We performed outlier filtering for hot (energetic) or lower pixels
in the data by following each pixel through time. This task was
performed in two steps, first flagging all pixels with intensity 8σ or
more away from the median value computed from the five preced-
ing and five following images. The values of these flagged pixels
were replaced with the local median value. In the second pass, we
flagged and replaced outliers above the 4σ level, following the same
procedure. The total fraction of corrected pixels was 0.26 per cent
for the 3.6 µm and 0.06 per cent for the 4.5 µm channel.
We estimated and subtracted the background flux from each im-
age of the time series. To do this, we performed an iterative 3σ
outlier clipping for each image to remove the pixels with values
associated with the stellar point spread function (PSF), background
stars or hot pixels, created a histogram from the remaining pixels
and fitted a Gaussian to determine the sky background.
We measured the position of the star on the detector in each im-
age incorporating the flux-weighted centroiding method
2
using the
background-subtracted pixels from each image for a circular region
with radius 3 pixels centred on the approximate position of the star.
While we could perform PSF centroiding using alternative methods
such as fitting a two-dimensional Gaussian function to the stellar
image, previous experiences with warm Spitzer photometry showed
that the flux-weighted centroiding method is either equivalent or su-
perior (Beerer et al. 2011;Knutsonetal.2012; Lewis et al. 2013).
The variation of the x and y positions of the PSF on the detectors
were measured to be 0.20 and 0.21 for the 3.6 µ m channel and 0.89
and 0.8 pixels for the 4.5 µm channel.
We extract photometric measurements from our data following
two methods. First, aperture photometry was performed with
IDL
routine APER using circular apertures ranging in radius from 1.5 to
3.5 pixels in increments of 0.1. We filtered the resulting light curves
for 5σ outliers with a width of 20 data points.
We also performed photometry with time-variable aperture with
a size scaled by the value of a quantity known as the noise pixel
parameter (Mighell 2005; Knutson et al. 2012), which is described
in section 2.2.2 of the IRAC instrument handbook and has been
used in previous exoplanet studies to improve the results of warm
Spitzer photometry (see e.g. Charbonneau et al. 2008; Knutson
et al. 2012; Todorov et al. 2013; O’Rourke et al. 2014). The noise
parameter depends on the full width at half-maximum of the stellar
PSF squared and is defined as
˜
β =
(
i
I
i
)
2
i
I
2
i
, (1)
where I
i
is the intensity detected by the ith pixel. We use each
image to measure the noise pixel parameter applying an aperture
with radius of 4 pixels, ensuring that each pixel is considered should
the border of the aperture cross that pixel. We extracted source fluxes
from both channels with the aperture radii following the relation
r =
˜
βa
0
+ a
1
, (2)
where a
0
and a
1
were varied in the ranges 0.8 to 1.2 and −0.4 to
+0.4 with step 0.1.
The best results from both photometric methods were identified
by examining both the residual root mean square (rms) after fitting
2
As implemented in the IDL box centroider.pro, provided on the Spitzer
homepage: http://irsa.ipac.caltech.edu.
the light curves from each channel as well as the white and red
noise components measured with the wavelet technique detailed in
Carter & Winn (2009). The second photometric method resulted in
the lowest white and random red noise correlated with data points
co-added in time or spectral sampling for the 3.6 µmdatawith
a
0
= 0.9 and a
1
=−0.1. For the 4.5 µm data, the first method gave
better results with aperture radius 2.4 pixels and sky annulus defined
between radii 2.40 and 6.48 pixels.
Finally, we also performed a visual inspection of the Spitzer
IRAC images for obvious stellar companions by aligning and stack-
ing ∼3000 of the available images in both channels, finding no ev-
idence for bright stellar sources within the 38 × 38 arcsec
2
field of
view.
2.3 Stellar variability monitoring
Stellar activity can complicate the interpretation of exoplanet trans-
mission spectra, especially when multi-instrument multi-epoch data
sets are combined (Pont et al. 2013). As the star rotates, active re-
gions on its surface enter into and out of view, causing the measured
flux to exhibit a quasi-periodic variability that introduces variation
of the measured transit depth (as R
p
/R
∗
). This becomes particularly
important when transit observations made over several months or
years are combined to construct an exoplanet transmission spec-
trum, as in our HST/Spitzer study.
We obtained high-resolution spectra of WASP-6 from the pub-
licly available HARPS data to search for evidence of chromospheric
activity. All spectra in the wavelength region of the Ca
II H&K
lines show evidence for emission, implying a moderate stellar ac-
tivity compared to the same region of HARPS spectra obtained
for HD189733. An examination of the HST and Spitzer white and
spectral light curves show no evidence for spot crossings.
We obtained nightly photometry of WASP-6 to monitor and char-
acterize the stellar activity over the past three observing seasons with
the Tennessee State University Celestron 14-inch (C14) automated
imaging telescope (AIT) located at Fairborn Observatory in Arizona
(Henry 1999). The AIT uses an SBIG STL-1001E CCD camera
and exposes through a Cousins R filter. Each nightly observation
of WASP-6 consists of 4–10 consecutive exposures that include
several comparison stars in the same field of view. The individual
nightly frames were co-added and reduced to differential magni-
tudes (i.e. WASP-6 minus the mean brightness of nine constant
comparison stars). Each nightly observation has been corrected for
bias, flat-fielding, pier-side offset and for differential atmospheric
extinction.
A total of 258 nightly observations (excluding a few isolated
transit observations) were acquired in five groups during three ob-
serving seasons between 2011 September and 2014 January. The
five groups are plotted in Fig. 3, where groups 2 through 5 have
been normalized to have the same mean magnitude as the first
group. The standard deviations of the five groups with respect to
their corresponding means are 0.0032, 0.0065, 0.0034, 0.0078 and
0.0053 mag, respectively. The precision of a single measurement in
good photometric conditions with C14 is typically 0.002–0.003 mag
(see, e.g. Sing et al. 2013). The standard deviations of groups 2, 4
and 5 significantly exceed the measurement precision and thus in-
dicate low-level activity due to star spots. A periodogram analysis
has been performed of each data set with trial frequencies ranging
between 0.005 and 0.95 c d
−1
, corresponding to a period range be-
tween 1 and 200 d. Period analysis of group 5 gives the clearest
indication of rotational modulation with a period of 23.6 ± 0.5 d
and a peak-to-peak amplitude of 0.01 mag. We take this period to
MNRAS 447, 463–478 (2015)
Downloaded from https://academic.oup.com/mnras/article-abstract/447/1/463/989512
by Queen's University of Belfast user
on 08 February 2018
