Transiting exoplanets from the CoRoT space mission⋆

 

 

A&A 555, A118 (2013)
DOI: 10.1051/0004-6361/201321462
c(cid:2) ESO 2013

Astronomy
&
Astrophysics

Transiting exoplanets from the CoRoT space mission(cid:2)

XXIV. CoRoT-25b and CoRoT-26b: two low-density giant planets

J. M. Almenara1, F. Bouchy1,2, P. Gaulme3, M. Deleuil1, M. Havel4, D. Gandolfi5, H. J. Deeg6,7, G. Wuchterl8,
T. Guillot9, B. Gardes1, T. Pasternacki10, S. Aigrain11, R. Alonso6,7,12, M. Auvergne13, A. Baglin13, A. S. Bonomo1,14,
P. Bordé15, J. Cabrera10, S. Carpano5, W. D. Cochran16, Sz. Csizmadia10, C. Damiani1, R. F. Diaz1, R. Dvorak17,
M. Endl16, A. Erikson10, S. Ferraz-Mello18, M. Fridlund5, G. Hébrard19,2, M. Gillon20, E. Guenther8, A. Hatzes8,
A. Léger15, H. Lammer21, P. J. MacQueen16, T. Mazeh22, C. Moutou1, M. Ollivier15, A. Ofir23, M. Pätzold24,
H. Parviainen6,7, D. Queloz12, H. Rauer10,25, D. Rouan13, A. Santerne1, B. Samuel13, J. Schneider26, L. Tal-Or22,
B. Tingley6,7,27, and J. Weingrill21

1 Aix Marseille Université, CNRS, LAM (Laboratoire d’Astrophysique de Marseille) UMR 7326, 13388 Marseille, France

e-mail: josemanuel.almenara@oamp.fr

2 Institut d’Astrophysique de Paris, 98bis boulevard Arago, 75014 Paris, France
3 Department of Astronomy, New Mexico State University, PO Box 30001, MSC 4500, Las Cruces, NM 88003-8001, USA
4 NASA Postdoctoral Program Fellow, Ames Research Center, PO Box 1, Moffett Field, CA 94035, USA
5 Research and Scientific Support Department, ESTEC/ESA, PO Box 299, 2200 AG Noordwijk, The Netherlands
6 Instituto de Astrofisica de Canarias, 38205 La Laguna, Tenerife, Spain
7 Universidad de La Laguna, Dept. de Astrofísica, 38200 La Laguna, Tenerife, Spain
8 Thüringer Landessternwarte, Sternwarte 5, Tautenburg 5, 07778 Tautenburg, Germany
9 Observatoire de la Côte d’Azur, Laboratoire Cassiopée, BP 4229, 06304 Nice Cedex 4, France
10 Institute of Planetary Research, German Aerospace Center, Rutherfordstrasse 2, 12489 Berlin, Germany
11 Department of Physics, Denys Wilkinson Building Keble Road, Oxford, OX1 3RH, UK
12 Observatoire de l’Université de Genève, 51 chemin des Maillettes, 1290 Sauverny, Switzerland
13 LESIA, Obs de Paris, Place J. Janssen, 92195 Meudon Cedex, France
14 INAF − Osservatorio Astronomico di Torino, via Osservatorio 20, 10025 Pino Torinese, Italy
15 Institut d’astrophysique spatiale, Université Paris-Sud 11 & CNRS, UMR 8617, Bât. 121, 91405 Orsay, France
16 McDonald Observatory, The University of Texas at Austin, Austin, TX 78712, USA
17 University of Vienna, Institute of Astronomy, Türkenschanzstr. 17, 1180 Vienna, Austria
18 IAG, Universidade de Sao Paulo, Brazil
19 Observatoire de Haute Provence, 04670 Saint Michel l’Observatoire, France
20 University of Liège, Allée du 6 août 17, Sart Tilman, Liège 1, Belgium
21 Space Research Institute, Austrian Academy of Science, Schmiedlstr. 6, 8042 Graz, Austria
22 Wise Observatory, Tel Aviv University, 69978 Tel Aviv, Israel
23 Institut für Astrophysik, Georg-August-Universität, Friedrich-Hund-Platz 1, 37077 Göttingen, Germany
24 Rheinisches Institut für Umweltforschung an der Universität zu Köln, Aachener Strasse 209, 50931 Köln, Germany
25 Center for Astronomy and Astrophysics, TU Berlin, Hardenbergstr. 36, 10623 Berlin, Germany
26 LUTH, Observatoire de Paris, CNRS, Université Paris Diderot, 5 place Jules Janssen, 92195 Meudon, France
27 Department of Physics and Astronomy, Aarhus University, 8000 Aarhus C, Denmark

Received 13 March 2013 / Accepted 24 May 2013

ABSTRACT

We report the discovery of two transiting exoplanets, CoRoT-25b and CoRoT-26b, both of low density, one of which is in the Saturn
mass-regime. For each star, ground-based complementary observations through optical photometry and radial velocity measurements
secured the planetary nature of the transiting body and allowed us to fully characterize them. For CoRoT-25b we found a planetary
−0.06 g cm−3. The planet orbits an F9 main-
mass of 0.27 ± 0.04 MJup, a radius of 1.08+0.3
sequence star in a 4.86-day period, that has a V magnitude of 15.0, solar metallicity, and an age of 4.5+1.8
−2.0-Gyr. CoRoT-26b orbits a
slightly evolved G5 star of 9.06 ± 1.5-Gyr age in a 4.20-day period that has solar metallicity and a V magnitude of 15.8. With a mass
−0.07 g cm−3, it belongs to the low-mass hot-Jupiter population.
of 0.52± 0.05 MJup, a radius of 1.26+0.13
Planetary evolution models allowed us to estimate a core mass of a few tens of Earth mass for the two planets with heavy-element
mass fractions of 0.52+0.08
−0.08, respectively, assuming that a small fraction of the incoming flux is dissipated at the center
of the planet. In addition, these models indicate that CoRoT-26b is anomalously large compared with what standard models could
account for, indicating that dissipation from stellar heating could cause this size.

−0.10 RJup and hence a mean density of 0.15+0.15

−0.07 RJup, and a mean density of 0.28+0.09

−0.15 and 0.26+0.05

Key words. planetary systems – techniques: photometric – techniques: radial velocities – techniques: spectroscopic

(cid:2) The CoRoT space mission, launched on December 27th 2006, has been developed and is operated by CNES, with the contribution of Austria,
Belgium, Brazil, ESA (RSSD and Science Programme), Germany and Spain. Partly based on observations obtained at the European Southern
Observatory at Paranal and La Silla, Chile in programs 083.C-0690(A), 184.C-0639.

Article published by EDP Sciences

A118, page 1 of 11

A&A 555, A118 (2013)

1. Introduction

More than 190 giant exoplanets (mass and radius larger
than 0.1 MJup, 0.5 RJup) transiting their parent stars have been
discovered, mostly with dedicated photometric surveys, such as
TrES (Alonso et al. 2004), WASP (Pollacco et al. 2006), and
HAT (Bakos et al. 2007) from the ground, and CoRoT and
Kepler from space (Baglin et al. 2009; Borucki et al. 2010).
Transiting planets constitute a unique testbed for precise mod-
eling of exoplanet structure and evolution, because the combina-
tion of photometric and spectrometric measurements can lead to
the accurate determination of the physical and orbital properties
of the planets and their host stars. This well-characterized planet
population allows one to probe their internal structure and fur-
thermore provides some hints on the physical processes that oc-
cur in their interior (e.g. Laughlin et al. 2011). While the number
of transiting planets in the Jupiter and higher mass regime has
increased to more than one hundred mostly well-characterized
members, planets in the Saturn mass domain (0.1−0.4 MJup) re-
main rare; there are currently only 18 with masses between 0.1
and 0.4 MJup and sizes larger than 0.5 RJup. We still need to better
assess how the properties of these two classes of the giant close-
in population compare and whether the same formation mecha-
nisms occur in both.

We report the discovery of CoRoT-25b and CoRoT-26b,
which belong to the low-mass range of the giant close-in pop-
ulation. In Sect. 2, we present the CoRoT observations and in
Sect. 3, the ground-based follow-up observations that were used
to confirm the planetary nature of the objects and determined
their masses. In Sect. 4, we describe the spectral analysis that
leads to the determination of the host stars’ fundamental parame-
ters, surface gravity, effective temperature, and abundances. The
orbital and physical parameters of the planets are then derived
with a global fitting method that combines photometric and ra-
dial velocity measurements (Sect. 5). The properties of these two
planets have been explored with two different approaches based
on evolution models, and the results are presented in Sect. 6. We
conclude in Sect. 7 with a brief comparison of these two planets
with the giant population.

2. CoRoT observations

The CoRoT observation run LRc02 corresponds to a 145-day
run from April 14 to September 7, 2008 in the Galactic center di-
rection. Three more planets have previously been reported from
this run: CoRoT-6b (Fridlund et al. 2010), CoRoT-9b (Deeg et al.
2010), and CoRoT-11b (Gandolfi et al. 2010).

The first transits of CoRoT-25b and CoRoT-26b were de-
tected 99 and 51 days after the begining of the run by the Alarm
Mode (Surace et al. 2008; Bonomo et al. 2012), in the monochro-
matic light curves (see Fig. 1) of the targets LRc02_E1_1280
and LRc02_E2_4747, respectively. After the initial detection,
the sampling rate of both targets was switched from 512 s to 32 s.
The transits of CoRoT-25b and CoRoT-26b were identified with
an orbital period of 4.9 and 4.2 days, and both have a tran-
sit depth of ≈0.5% (contaminated values). Coordinates, iden-
tifications, magnitudes, and additional parameters are given in
Table 3.

Factors for photometric contamination of the targets’ flux
from nearby stars were computed for each CoRoT aperture mask
(Fig. 2) with two methods. We first used the approach devel-
oped by Pasternacki et al. (2011), which consists of modeling
the point spread function (PSF) of the target and contaminating
stars and of calculating its flux contribution in the photometric

A118, page 2 of 11

aperture. Contamination factors of 1.3 ± 0.3% and 10.9 ± 0.9%
for CoRoT-25 and CoRoT-26 are obtained. Second, we ap-
plied the method from Gardes et al. (2011), which is concep-
tually similar to the previous one, but found contaminations
of 3.4 ± 0.9% and 10.4 ± 1.2%. Results from the two methods
agree for CoRoT-26, but differ by two sigmas for CoRoT-25.
The difference in the contamination estimates arises from possi-
ble differences in recentering the stellar PSF model on the photo-
metric mask, because CoRoT does not provide individual images
for each exposure. Finally, we decided to adopt weighted mean
values from both methods, that is, 1.8 ± 0.3% for CoRoT-25
and 10.7 ± 0.7% for CoRoT-26. In both cases, the errors of the
contamination factors (which affect the relative depth of the tran-
sits) are too small to contribute significantly to the final system
parameter errors reported in Table 3.

CoRoT light curves were processed with the CoRoT Data
Analysis pipeline, which removes signatures in the light curves
that are correlated with systematic error sources from the tele-
scope and spacecraft, such as pointing drift, focus change, and
flag features due to high-energy-particle impact, and thermal
transients (Drummond et al. 2006; Auvergne 2006; Drummond
et al. 2008). The remaining jumps and discontinuities were cor-
rected. For the analysis, we kept only the light curves covering a
span of 2.4 h for CoRoT-25 and 8.2 h for CoRoT-26 before and
after each transit, corrected for the contamination, and binned
the data sampled with the 32 s rate those sampled with 512 s.
In addition, we removed some remaining low-frequency modu-
lations by subtracting a parabolic fit of the off-transit data and
normalized the light curves, as has generally been performed for
CoRoT planet detections (e.g. Alonso et al. 2008).

3. Ground-based follow-up observations
3.1. Photometricmeasurements

The main objective of ground-based photometric follow-up is
to check whether the observed transit features occur on the
target star and not on a potential background eclipsing-binary
system (Deeg et al. 2009). Following-up CoRoT exoplane-
tary candidates is challenging because of the faintness of the
targets and the need for time-critical observations, given the
transit ephemeris. For this reason, the photometric follow-up
of the candidate CoRoT-25b was performed with three tele-
scopes. Observations with short on- and off-transit coverages
were acquired on June 6, 2010 with the Canada France Hawaii
Telescope (CFHT), and on June 11, 2010 with the 1.2-m Euler
telescope at La Silla Observatory. Both indicated a detection
of the transit on the target star, but because of timing errors
from ephemeris uncertainties, we did not consider this result as
sufficiently reliable. Longer coverage of a full transit was ob-
tained on June 15, 2010 with the IAC80 telescope at the Teide
Observatory, which confirmed that the transits occur on the tar-
get star, with a depth of about 0.5%. Nearby eclipsing binaries,
at distances larger than ∼1.5(cid:5)(cid:5) from the target, could therefore be
excluded as a source of a false positive.

Photometric follow-up observations of CoRoT-26b were per-
formed in July and August 2008 with both the IAC80 and 1.2-
m Euler telescopes. The runs consisted of short on-off observa-
tions in and out of transits. The observations from the IAC80
were unable to detect the transit in the target star, but their qual-
ity in combination with a smaller timing error was sufficient to
exclude that CoRoT’s signal could arise from any nearby con-
taminator. Then, the Euler observations showed a clear transit
on the target with a depth of 0.8 ± 0.3%, fully compatible with
the photometric dimming measured by CoRoT and rejecting that

J. M. Almenara et al.: CoRoT-25b and CoRoT-26b

Fig. 1. CoRoT light curve from raw data of CoRoT-25 (top) and CoRoT-26 (bottom). The changes in sampling rates from 512 to 32 s appear
as increases in the scatter. In gray dots, the 32 s integrations are binned to 512 s. The light curves show discontinuities due to the impact of
high-energy particles and other instrumental effects. The transits can be seen as regular dips.

2’x2′ POSS image centered on Corot-25

N

E

RA – 280.629700 (arcsec)

2’x2′ POSS image centered on Corot-26

13
8

14

2

7

1

5

6

9
1718

16

4
15
1011
2021

3

12
19

14

13

8

4

7

11

2

6

1

9

3
5
10

12

#: Rmag

0: 14.6
1: 19.6
2: 20.3
3: 20.0
4: 18.2
5: 19.8
6: 19.0
7: 16.9
8: 19.7
9: 19.1
10: 19.7
11: 19.9
12: 20.6
13: 19.2
14: 19.6
15: 18.5
16: 20.3
17: 19.5
18: 19.1
19: 16.6
20: 20.6

#: Rmag

0: 15.3
1: 18.5
2: 16.2
3: 19.6
4: 18.3
5: 18.4
6: 17.2
7: 17.4
8: 18.4
9: 17.9
10: 16.5
11: 17.9
12: 19.0
13: 17.1
14: 19.4

)
c
e
s
c
r
a
(
0
0
9
3
1
5

.

6
+
E
D

)
c
e
s
c
r
a
(

0
2
9
9
6
9
6
+
E
D

.

60

0

20

0

20

0

60

0

0

20

0

20

0

0

N

E

RA – 279.750550 (arcsec)

Fig. 2. POSS image of CoRoT-25 and CoRoT-26 (marked with a red
cross in the center of the field) with the CoRoT mask superimposed (in
blue solid line), and the contaminating stars (red crosses identified with
a number, the contaminating R-magnitude its at the right of the image).

any eclipsing binary at distances larger than ∼1.5(cid:5)(cid:5) from the tar-
get may have caused a false alarm.

3.2. Spectroscopic and radial velocity measurements

Radial velocity (RV) measurements were performed with the
HARPS spectrometer at the focus of ESO’s 3.6-m telescope

at La Silla, as part of the ESO large program 184.C-0639.
Some observations were also obtained with HIRES on the 10-m
Keck-1 telescope, as part of a NASA key science project to sup-
port the CoRoT mission. HARPS was used with the observ-
ing mode obj_AB, without simultaneous thorium calibration to
monitor the Moon background light on its second fiber. The
exposure time was set to one hour. We reduced the HARPS
data and computed the RVs with a pipeline based on the cross-
correlation techniques (Baranne et al. 1996; Pepe et al. 2002).
Radial velocities were obtained by weighted cross-correlation
with a numerical G2 mask. For some exposures contaminated
by moonlight, we made the correction described by Bonomo
et al. (2010), which consist of substracting the cross-correlation
function (CCF) from the second fibre, which contains the Sun
spectrum (reflected by the Moon), from the stellar CCF. HIRES
observations were performed with the red cross-disperser and
the I2-cell to measure RVs. We used the 0.861-arcsec wide slit
that leads to a resolving power of R ≈ 45 000. Differential radial
velocities were computed using the Austral Doppler code (Endl
et al. 2000).

3.2.1. Radial velocities of CoRoT-25

A set of 28 spectra was recorded for CoRoT-25 with HARPS
between September 9, 2009 and September 22, 2011, includ-
ing nine consecutive exposures of twice one hour made during
one night. The signal-to-noise ratio (S/N) per pixel at 550 nm
ranges from 5 to 14. Five measurements were slightly affected
by the moonlight. The correction provided by the substraction
of the Moon CCF from the stellar CCF induced an RV change
of up to 76 m s−1 for these five measurements. A set of six ex-
posures was recorded with HIRES on June 29−30, 2009 and on
July 23−24, 2011, including two consecutive exposures of twice
one hour made during one night without Moon contamination.

The HARPS and HIRES radial velocities are given in
Table 1. The two sets of relative radial velocities were simultane-
ously fitted with a Keplerian model, where the transit epoch and
period are fixed at values from a first modeling of the CoRoT
light curve, and where an offset was adjusted between the two
different instruments. No significant eccentricity was found, and
we decided hereafter to set it to zero. The best solution is ob-
tained for a semi-amplitude K = 30.0 ± 4.6 m s−1 and an offset
for the HIRES radial velocities of −15.4185 ± 0.007 km s−1. The

A118, page 3 of 11

Fig. 3. Phase-folded radial velocities of CoRoT-25. The black circles
and green squares correspond to HARPS and HIRES measurements.
The red open circles correspond to the averages of the double (same
night) measurements.

Table 1. Radial velocity measurements of CoRoT-25 obtained with
HARPS and HIRES.

A&A 555, A118 (2013)

BJD
−2 400 000

RV
[km s−1]

±1σ
[km s−1]

HARPS 3.6-m ESO

55 021.70233
55 022.67772
55 024.73260
55 340.82475
55 341.86346
55 342.84784
55 352.69145
5 5354.77688
55 680.82312
55 680.86341
55 682.83514
55 682.87834
55 685.81512
55 694.82523
55 694.86761
55 695.82976
55 695.87255
55 712.79965
55 712.84286
55 777.61927
55 777.66206
55 802.55826
55 802.60376
55 804.54140
55 804.58589
55 823.51409
55 826.51248
55 826.55967

−15.3885
−15.3792
−15.3885
−15.4012
−15.4267
−15.4039
−15.3524
−15.4629
−15.4284
−15.4705
−15.3846
−15.3889
−15.4460
−15.4400
−15.4321
−15.4756
−15.4268
−15.3845
−15.3852
−15.4456
−15.4633
−15.4837
−15.4103
−15.4077
−15.4131
−15.3407
−15.4445
−15.4884
HIRES 10-m Keck

55 012.823144
55 013.076915
55 013.958030
55 014.085841
55 766.967095
55 767.997414

0.01851
−0.00302
0.00818
−0.02283
−0.00852
0.00769

0.0274
0.0587
0.0366
0.0311
0.0207
0.0506
0.0313
0.0260
0.0358
0.0344
0.0192
0.0197
0.0167
0.0156
0.0149
0.0181
0.0276
0.0200
0.0171
0.0335
0.0319
0.0402
0.0351
0.0456
0.0570
0.0332
0.0178
0.0146

0.0117
0.0178
0.0168
0.0152
0.0150
0.0153

Notes. BJD is the barycentric Julian date.

dispersion of the residuals is 25 m s−1 and the reduced χ2 is 1.1.
We also computed the average of each double measurement (ob-
tained during the same night) and found a very similar result with
K = 31.6 ± 3.9 m s−1 and a residual dispersion of 10 m s−1. The
joint analysis of the photometric and RV data in Sect. 5 does
not change these results. Figure 3 shows all RV measuremens
after subtracting the RV offset and a phase-folding to the orbital
period.

To examine the possibility that the RV variation is due to a
blended-binary scenario – a single star plus an unresolved eclips-
ing binary –, we followed the procedure described in Bouchy
et al. (2009b). It consists of checking the spectral line asymme-
tries and the dependance of the RV variations as a function of
the cross-correlation mask (a template spectrum made of box-
shaped emission lines at the position of selected spectral lines for
a specific spectral type − see Baranne et al. 1996). To increase
the S/N of the cross-correlation function (CCF) bisectors, we av-
eraged the CCFs close to the same orbital phase φ, at φ = 0.25
and φ = 0.75. Sixteen and 11 CCFs were averaged with or-
bital phases in the ranges of 0.1−0.4 and 0.6−0.9, respectively.
The bisectors were computed on these two averaged CCFs; no
significant variation of the bisector span was observed (Fig. 4).

A118, page 4 of 11

Fig. 4. HARPS bisector span versus radial velocities of CoRoT-25. The
two red circles correspond to the averaged cross-correlation functions
at the extreme phases of φ = 0.25 and 0.75.

Furthermore, the radial velocities were computed with different
cross-correlation templates (F0, G2, and K5) without a signif-
icant change in the amplitude of the RV variations. These two
checks exclude some configurations of blended binaries, which
supports the planetary nature of CoRoT-25b.

No significant radial velocity drift (<100 m/s) was de- tected during the two years covered by the HARPS and HIRES measurements. This excludes the presence of any other massive giant planet (>5 MJup) at closer than 3 AU in the system.

3.2.2. Radial velocities of CoRoT-26

A set of 27 spectra was recorded for CoRoT-26 with HARPS
between July 10, 2009 and September 23, 2011, including seven
consecutive exposures of twice one hour made during one night
(Table 2). Three measurements with a S/N (per pixel at 550 nm)
lower than 4 were rejected. The remaining 24 measurements,
listed in Table 2, have an S/N from 4 to 8. The first measurement
was slightly affected by moonlight, but the RV of the Moon was
far away from the stellar RV and did not require correction. The
fourth measurement was affected by the Moon, with its RV close
to the stellar one. The correction induced a change of 225 m s−1
for this exposure. All other measurements were made during

Table 2. Radial velocity measurements of CoRoT-26 obtained with
HARPS.

J. M. Almenara et al.: CoRoT-25b and CoRoT-26b

BJD
−2 400 000
55 022.63404
55 024.68691
55 064.56484
55 320.90243
55 323.81802
55 324.85402
55 334.81533
55 683.88991
55 683.91624
55 685.87092
55 685.90796
55 686.86764
55 686.90771
55 692.91083
55 693.82839
55 693.87075
55 716.80319
55 779.61691
55 779.65885
55 806.53474
55 825.50887
55 825.55249
55 827.51534
55 827.56043

RV
[km s−1]
15.5780
15.4251
15.5480
15.5432
15.5348
15.5795
15.4603
15.5118
15.4448
15.5648
15.5332
15.5711
15.5169
15.4867
15.4613
15.4756
15.6194
15.5863
15.5923
15.4142
15.5852
15.5958
15.4103
15.5039

±1σ
[km s−1]
0.0372
0.0326
0.0337
0.0407
0.0256
0.0448
0.0515
0.0423
0.0367
0.0302
0.0382
0.0325
0.0302
0.0226
0.0387
0.0300
0.0521
0.0398
0.0526
0.0452
0.0296
0.0274
0.0473
0.0489

Fig. 5. Phase folded radial velocities of CoRoT-26. Symbols are similar
to Fig. 3.

Notes. BJD is the barycentric Julian date.

dark time. We fitted the dataset with a Keplerian model, where
the transit epoch and period are fixed at values from a first mod-
eling of the CoRoT light curve. No significant eccentricity was
detected, and hereafter it was set it to zero. The best solution is
obtained for a semi-amplitude K = 56.9 ± 8.4 m s−1. The disper-
sion of the residuals is 35 m s−1 and the reduced χ2 is 0.98. We
also computed the mean of each double measurement (obtained
during the same night) and obtained K = 64.5 ± 11 m s−1 com-
patible within 1σ to the previous determination and a residual
dispersion of 20 m s−1. The posterior joint analysis of the pho-
tometric and RV data did not change these results within error
bars. In Fig. 5, we plot the radial velocity measurements phase-
folded to the orbital period.

To examine whether the radial velocity variation could be
caused by a diluted eclipsing binary, we averaged the CCFs close
to the same extreme phases. Five and seven CCFs were averaged
at orbital phases in the ranges 0.2−0.3 and 0.7−0.8. Bisector
spans were computed on these two averaged CCFs, but showed
no significant variations (Fig. 6). Furthermore, the radial veloc-
ities were computed with different cross-correlation templates
(F0, G2 and K5) without a significant change in the amplitude
of the RV variations. These two checks exclude some configura-
tions of a blended binary, which supports the planetary nature of
CoRoT-26b.

No significant RV drift (<200 m/s) was found over the two years of HARPS data, excluding any additional massive sub- stellar companion (>10 MJup) at closer than 3 AU in the system.

4. Stellar parameters and interstellar extinction
We obtained a spectrum of CoRoT-25 with UVES/VLT using
the 390 + 580 nm setting, which covers the wavelength range
from 327.4 to 450.6 nm and 478.5 to 681.7 nm. With the

Fig. 6. HARPS bisector span versus radial velocities of CoRoT-26.
Symbols are similar to Fig. 4.

slit-width of 0.8 arcsec, the spectrum has a resolution λ/Δλ ∼
50 000. Three spectra were taken in service mode (program
083.C-0690(A)) on September 25, 2010, September 28, 2010,
and September 29, 2010, each exposed for one hour. We used
IRAF routines to remove the bias offset, to flat-field the data, to
remove cosmic ray hits, and to extract and wavelength-calibrate
the spectrum. The three spectra, set in the rest frame, were then
co-added and used to derive a first estimate of the spectral type
of the host star and to confirm that it is a dwarf-type star. We
then took advantage of the HARPS spectra that were collected
for the RV analysis, because they provide a higher spectral reso-
lution. These spectra, once set in the rest frame, were co-added
following the methodology described for previous CoRoT plan-
ets (e.g. Rauer et al. 2009) and produced spectra with a typical
S/N of 230 for CoRoT-25 and 130 for CoRoT-26 at 5600 Å in the
continuum. These spectra were analyzed to derive the effective
temperature, surface gravity, and abundances of some elements.
As described in Fridlund et al. (2010), to that purpose we used
several methods: Balmer-line fitting and the SME and VWA
packages. For the surface gravity, we used the lines of Mg ib
at 5184 Å and of Ca i at 6122 Å, 6162 Å and 6439 Å as diagnos-
tics. Values obtained for effective temperature, surface gravity,
metallicity, micro- and macro turbulence, and v sin i are given in
Table 3. The latter was estimated by selecting a set of isolated

A118, page 5 of 11

A&A 555, A118 (2013)

and unblended metal spectral lines. We also checked for consis-
tency with the v sin i estimate from the HARPS CCF.

The fundamental parameters of the stars were finally de-
rived by comparing the position of the stars in the HR diagram
in the Teff, M1/3
/R(cid:2) plane with stellar evolutionary tracks from
(cid:2)
STAREVOL (Palacios, priv. comm.). Using the spectroscopic
values and the M1/3
/R(cid:2) obtained from the transit fitting and their
(cid:2)
associated error bars, we generated a series of Gaussian ran-
dom realizations of Teff, [Fe/H], and M1/3
/R(cid:2). For each realiza-
(cid:2)
tion, we determined the best evolutionary track using a χ2 min-
imization on these three parameters. The given errors on the
parameters account only for the statistical errors. The errors due
to the models are certainly higher, but it is not clear by how
much. We abstained therefore from any increase of these er-
rors, which would have been arbitrary. We found CoRoT-25 to
be an F9 dwarf star, and CoRoT-26 is a slightly evolved G5 star
(see Table 3). Note that the surface gravities inferred from these
stellar masses and radii (log g = 4.31+0.05
−0.08 for CoRoT-25 and
log g = 3.99+0.05
−0.08 for CoRoT-26) are consistent with the spec-
trometric estimates (log g = 4.28 ± 0.10 for CoRoT-25, and
log g = 4.10 ± 0.10 for CoRoT-26).

The inferred evolutionary ages are 4.5+1.8

−2.0 Gyr and 9.06 ±
1.5 Gyr for CoRoT-25 and CoRoT-26, respectively. These esti-
mates, which point toward evolved main-sequence systems, are
consistent with a lack of activity and slow rotation velocities of
the stars. None of the stellar spectra display any evidence of
activity, thus we could not derive chromospheric ages for the
stars. Both stars also have a low v sin i. By assuming the sim-
plest configuration, where the stellar spin axis is perpendicular
to the line of sight, such v sin i would correspond to rotation
periods of 14.1 ± 2.9 days and 45.3 ± 25 days for CoRoT-25
and CoRoT-26. No significant peaks were found in the peri-
odogram of the light curve. The age deduced from gyro chronol-
ogy (Barnes 2007; Mamajek & Hillenbrand 2008) is 2.2 Gyr for
CoRoT-25 and 9.2 Gyr for CoRoT-26. These two values agree
well with the ages deduced from evolutionary tracks.

The Li i line at 6708 Å is present in both spectra with an
equivalent width of 70.0 mÅ in CoRoT-25 and 65.0 mÅ for
CoRoT-26. While lithium has often been considered as an age
indicator because the depletion of lithium appears to be de-
termined by the convection zone depth, recent analyses (e.g.
Takeda et al. 2007; Israelian et al. 2009) have shown that the
dispersion in Li abundances at a given Teff is high, weakening
the dependance of Lithium abundance upon age. This disagree-
ment between age indicators, namely Li i abundance on the one
hand and activity indicators and the stars’ rotation on the other
hand, has previously been noticed in other planet host-stars (see
Cabrera et al. 2010).

Interstellar extinctions (Av) and distances (d) to the tar-
get stars were derived by using the spectral energy distribution
(SED) fitting procedure described in Gandolfi et al. (2008). For
this purpose, we merged the ExoCat optical photometry within
ExoDat (Deleuil et al. 2009) with the 2MASS (Cutri et al. 2003)
and WISE (Wright et al. 2010) infrared data. We then assumed
both stars to emit as black bodies, with the above estimated
effective temperatures and radii, to find that Av = 0.70±0.07 mag
and d = 1000+85
−25 pc for CoRoT-25, and Av = 0.85 ± 0.10 mag
and d = 1670+205
−110 pc for CoRoT-26.

Figure 7 shows the de-reddened SED of the target stars. We
also overplotted the synthetic stellar spectra (light-blue lines) ob-
tained with the NextGen model (Hauschildt et al. 1999), by us-
ing the photospheric parameters of CoRoT-25 and CoRoT-26.

A118, page 6 of 11

Fig. 7. Optical and infrared photometric measurements (circles) and
fitted spectral energy distributions (blue line) of CoRoT-25 (top) and
CoRoT-26 (bottom).

The WISE data at 12 and 22 μm are upper limits and were not
included in the fitting procedure.

5. Bayesian modeling with MCMC

We aim to estimate the orbital parameters (semi-major axis a,
epoch of periastron t0, orbital period P, eccentricity e, argu-
ment of periastron ω), the inclination angle i, the stellar limb-
darkening u1, u2 (of the quadratic approximation), and the phys-
ical properties of the planetary companion (mass Mp, radius Rp).
The main difficulty in modeling light curves of planetary
transits arises from the strong coupling between some parame-
ters, in particular inclination, semi-major axis, and ratio of plan-
etary to stellar radii. Thus, a joint analysis of radial velocity and
photometric measurements with a Bayesian approach is more
and more often used by modelers (e.g. Gillon et al. 2009, Pont
et al. 2009). The advantage of joint modeling is the tightening
of constraints on parameters that can be extracted from both
datasets (P, t0, e, ω). The interest of Bayesian modeling lays in
maximizing the posterior probability P(M|D, I) of a model given
some prior informations, instead of maximizing the likelihood
L ≡ P(D|M, I) of the dataset with respect to a model (i.e., least-
squares method for a normal noise distribution).

Because we considered one model type, we used the sim-
plest way of applying the Bayesian approach, i.e., the maximum
a posteriori estimator (MAP), where we maximized LMAP =
P(M|I) P(D|M, I). In practice, it is equivalent and easier to min-
imize the quantity lMAP = − log LMAP. Therefore, because priors
are often defined as Gaussian functions centered on the expected

Table 3. CoRoT-25 and CoRoT-26 system parameters.

J. M. Almenara et al.: CoRoT-25b and CoRoT-26b

Name
CoRoT target ID
CoRoT ID
USNO-A2 ID
2MASS ID

Coordinates

Magnitudes

RA (J2000) [hh:mm:ss.ss]
Dec (J2000) [dd:mm:ss.ss]

U RGO
B Harris
V Harris
r(cid:5) Sloan-Gunn
i(cid:5) Sloan-Gunn
J 2MASS
H 2MASS
Ks 2MASS
WISE W1
WISE W2
WISE W3
WISE W4

Results from radial velocity observations

Orbital eccentricity e
Radial velocity semi-amplitude K [ m s−1]
Systemic velocity Vr [ km s−1]

Results from combined light curve and radial velocity analysis

Planet orbital period P [days]
Epoch of conjuntion tc [HJD]
Transit duration W [h]
Radius ratio r = Rp/R∗
Impact parameter b
Linear limb-darkening coefficient u1
Systemic velocity Vr [ km s−1]
1 − e2P1/3 [km s−2/3]
K2 = K

√

Deduced parameters

∗ /R∗ [solar units]

Scaled semi-major axis a/R∗
M1/3
Inclination i [deg]
Radial velocity semi-amplitude K [ m s−1]

Stellar parameters from spectroscopy
Effective temperature Teff [K]
Surface gravity log g [cgs]
Metallicity [Fe/H] [dex]
Rotational velocity v sin i [ km s−1]
Microturbulent velocity vmicro [ km s−1]
Macroturbulent velocity vmacro [ km s−1]

Star mass [M(cid:9)]
Star radius [R(cid:9)]
Deduced stellar surface gravity log g [cgs]
Age of the star t [Gyr]
Distance of the system [pc]
Extinction AV [mag]
Orbital semi-major axis a [AU]
Planet mass Mp [MJ]
Planet radius Rp [RJ]
Planet surface gravity log gp [cgs]
Planet mean density ρp [g cm−3]
Planet mean density ρp [ρJ]
Equilibrium tem