Possible Detection of Phase Changes from the Non-Transiting Planet HD

Home , HD 46375

arXiv:1011.2690v1 [astro-ph.EP] 11 Nov 2010 0 U naqaicrua orbit quasi-circular a on AU, 5 h trH 67 sarltvl rgt( bright relatively a is 46375 HD star The Introduction 1. naCDnral eiae oteatrsimlg pro- asteroseismology photom the ultra-precise an to obtain to dedicated 2002), al. normally et observed (Baglin and known gram CCD program such star a additional a As CoRoT with view. host the in of star by its field targeted brightest accessible was transit CoRoT the it the not is in does planet 46375 close-in planet HD 2000). The (Henry 2006). al. et Butler cec rgaso S,EAsRS,Asra egu,Braz of Belgium, participation Austria, with RSSD, ESA’s CNES, Spain. ESA, and the Germany of by Programs operated Science is and developed edo Send . . uy2,2018 25, July ⋆ Astrophysics & Astronomy 226(19) 4 0sa,kont otaStr-aspae ( planet Saturn-mass a host to known star, K0 04) .J Deeg J. H. h oo pc iso,luce n20 eebr2,was 27, December 2006 on launched mission, space CoRoT The .Gaulme P. ohs o-rniigStr-aseolntwt 3.023 a 2008. with of exoplanet fall Saturn-mass the non-transiting a host to T t eaieapiueis amplitude relative Its e words. Key s exoplanetary between overlap Plato. as the such highlights comb 2010a) The Kepler. al. and et CoRoT non-transit by that observations demonstrates present-day it with case, any In field. same the Conclusions. cur light same the Results. of analysis seismic a Methods. of means by model star Aims. Context. elrcsga iha1dypro.W hwta hssga i signal this that show We period. day 1 a with signal telluric 1 eoiysga ya ot2 most at by signal velocity 2 rpitoln eso:Jl 5 2018 25, July version: online Preprint e-mail: 3 4 5 rmtesm oo u sareference. a as run CoRoT same the from 8 6 9 10 osbedtcino hs hne rmtenon-transitin the from changes phase of detection Possible ff , 7 b ntttdAtohsqeSail,UR81,Université 8617, UMR Spatiale, d’Astrophysique Institut aoaor ieu nvri´ eNc,CNRS-Observatoi Nice, Université de Fizeau, Laboratoire aoaor asoé nvri´ eNc,CNRS-Observ Nice, Cassiopée, Université de Laboratoire EI,UR80,Osraor ePrs 29 ednCedex, Meudon 92195 Paris, de Observatoire 8109, UMR LESIA, nvriyo ina nto srnm,Türkenschanzstr Astronomy, of Inst Vienna, of University xodAtohsc,Uiest fOfr,Ofr X 3RH, OX1 Oxford Oxford, of University Astrophysics, Oxford UH M 12 bevtied ai,915Muo ee,F Cedex, Meudon 92195 Paris, de Observatoire 8102, UMR LUTH, rn eussto requests print ATUR57,CR n nvriyP aair 4A.E Bel E. Av. 14 Sabatier, P. University and CNRS 5572, LATT-UMR nttt eAto´sc eCnra n nvria eLa de Universidad and Canarias Astrof´ısica de de Instituto ≃ M [1880 eatmtt dniya pia aeegh,techanging the wavelengths, optical at identify to attempt We [email protected] Jup efideiec fasnsia inlcmail ntrsof terms in compatible signal sinusoidal a of evidence find We 6 h rsn okdaswt h eeto fpaecagsin changes phase of detection the with deals work present The h aaaayi eiso h ore pcrmadtefold the and spectrum Fourier the on relies analysis data The , 7 riigi .2536)dy,a 0 at days, 3.023573(65) in orbiting ) 1 , .Régulo C. , .Vannier M. , 00 yasmn radius a assuming by K 2030] tr:paeaysses ehius htmti,Metho photometric, Techniques: systems, planetary Stars: hsdtcino eetdlgtfo o-rniigplan non-transiting a from light reflected of detection This .Auvergne M. .Gaulme P. : ∆ F 6 aucitn.gaulme˙HD46375b no. manuscript e , p 2 7 / σ .Guillot T. , .Aigrain S. , = F RV ⋆ 0 oee,tetn lntr inli togybeddb a by blended strongly is signal planetary tiny the However, . = . 6(6 Mrye l 2000, al. et (Marcy 063(26) 4 lntH 67bb CoRoT by 46375b HD planet [13 .Baglin A. , . 0 , 26 m 3 8 .Mosser B. , V . R .Schneider J. , ]pm mliga albedo an implying ppm, 8] = = 4 1 7 .Baudin F. , . 1 . 89, R M . Jup 0398(23) p M sin n ninclination an and 4 n lnt a ecaatrzduigutapeiephoto ultra-precise using characterized be can planets ing .Mary D. , bol ai u,945OSYCedex ORSAY 91405 Sud, Paris 7 T18 ina Austria Vienna, 1180 AT 17, . upesd u o lmntd hnuigtetm series time the using when eliminated, not but suppressed, s tied aCoedAu,034Nc ee ,France 4, Cedex Nice 06304 Côte d’Azur, la de atoire etric ABSTRACT ed aCoedAu,018Nc ee ,France 2, Cedex Nice 06108 Côte d’Azur, la de re a eid twsosre yteCRTstliefr3 day 34 for satellite CoRoT the by observed was It period. day 6 ineadatrsimlg n hw h ihptnilof potential high the shows and asteroseismology and cience 9 i the .Bruntt H. , aua 80,L aua eeie Spain Tenerife, Laguna, La 38205, Laguna, il, = = eadteueo rudbsdsetooaiercobserva spectropolarimetric ground-based of use the and ve 1 nddtcino oa-yeoclain ntesm targe same the on oscillations solar-type of detection ined , .Catala C. , UK hsso h lnta tobt t tr ete r oimpro to try then We star. its orbits it as planet the of phases 10pm ntems aoal ae n a civdi only in achieved was and cases favorable most the in ppm) (100 n eas ftefiteso h inl fteodro 10 of order the of signal, the of faintness the of because ing 2002). Guillot & Showman (e.g. circulation spheric rapaesitmyb u ohrzna aitosi h clo the in variations modulatio horizontal to the (e.g due and of be atmospheres coverage may asymmetry the shift phase the in a Finally, or particles 2000). small al. b et of and si Seager scattering presence a non-isotropic the by from caused tray Departures be th may structure. of shape their evolution soidal thermal therefore the and direc of has proper planets calculation This atmospheric the albedo. planetary for the the sequences constrain particular to in amplitu and us the ties, reasons: helps several modulation for the interesting very is sphere brightnes the 2003). on Vannier & limits (Guillot upper planet place the or detect and lightcurve France n 10 olue France Toulouse, 31400 in, rance 2 n ftetm series. time the of ing s aaaayi,Sas niiul D46375 HD individual: Stars: analysis, data ds: .W Weiss W. W. , tsol ecnral ihalne oo bevto of observation CoRoT longer a with confirmable be should et nteotcldmi,ti eeto seteeychalleng extremely is detection this domain, optical the In h eeto frflce ih rma xpaeayatmo- exoplanetary an from light reflected of detection The neolntr ytm D435i oa nlgknown analog solar a is 46375 HD system. exoplanetary an ohapiueadpaewt ih eetdb h planet. the by reflected light with phase and amplitude both A 4 = .Deheuvels S. , i 4 [0 = .Michel E. , . 45 16 ◦ , / ramshrctmeaue n rc h atmo- the trace and temperatures atmospheric or t ria hs di phase orbital Its . 0 . 3 radyievsbebihns temperature brightness visible dayside a or 33] 5 .X Schmider F.-X. , 4 4 n .Samadi R. and , ohrsga,wihw trbt oa to attribute we which signal, nother .F Donati J.-F. , ⋆ ff r rmta fteradial- the of that from ers 2 10 .Bourguignon S. , .Appourchaux T. , 4 erclightcurves metric o D46179 HD for s(Gaulme ts tions. mission a

c during s S 2018 ESO ethe ve g 3 con- t eof de , 1 of s nu- , ese ud − e- ., n a 4 - - 2 Gaulme et al.: reﬂected light from HD 46375b few cases. Sing & L´opez-Morales (2009) detected the secondary 400 transit of OGLE-TR-56b in the z’-band, a case for which the Data ) Background fitting 1 planet is so hot that thermal emission from the planet contributes − 300 to the 363 91 ppm signal. Using CoRoT data, Snellen et al. Hz µ

(2009) identified± the changing phases of the planet CoRoT-1b 2 200 (with an amplitude 130 ppm) and its secondary transit was ∼ 100 independently measured by Alonso et al. (2009b). Secondary P (ppm transits of CoRoT-2b (a 60 20 ppm signal) were detected by ± Alonso et al. (2009a). With Kepler data, Borucki et al. (2009) 0 0.2 0.4 0.6 0.8 1 1.2 measured the secondary eclipse of HAT-P-7b (130 11 ppm). ν (cycles per day) With HD 46375, we have the opportunity to the detect± planetary phase changes of a non-transiting exoplanet. Such a kind Fig. 1. Power density spectrum calculated with DFT, with an detection was achieved in the infrared with the Spitzer Space oversampling factor of 8. The dot-dashed curve indicates the Telescope for υ And (Harrington et al. 2006) (amplitude 2900 mean stellar activity. The grey region indicates the range in ppm) but it would be the first time for optical wavelengths.∼ which we expect to detect the planetary signal. In a separate paper (Gaulme et al. 2010a, A&A, submit- ted), we used spectroscopic data and seismic analysis of the CoRoT lightcurves to constrain the fundamental parameters of the relative amplitude ∆F /F of the expected signal is in the HD 46375.Thestar may be consideredas a solar analog,of mass p ⋆ range [0.8, 61] ppm. Such relative fluctuations of intensity have 0.97 0.05 M ,age2.6 0.8 Gyr, metallicity [Fe/H] = 0.39 0.06 to be compared with the photon noise. In a single channel, dex,± and temperature⊙ T± = 5300 60 K. The stellar rotation± is eff the amount of photoelectrons collected by the CCD is about expected to be in the range [37, 51]± days. The spectrometric and N = 1.3 109 10 0.4mV s 1. Hence, for a m = 7.9 star, the asteroseismic results provide an estimate of the inclination of ⋆ − − V relative photon× noise× is 1/ √N = 10 3 per second. If individ- the stellar rotation axis i = 50 18 and the planetary mass ⋆ − ◦ ual measurements are binned in 6 samples per orbital period, the M sin i = 0.234 0.008∗ M . By± assuming that planet orbits p Jup noise level reaches 1.5 ppm for each orbital sample, for a time in the stellar equatorial± plane (see however Hebrard et al. 2010), coverage of 11 orbits (i.e., 33 days). Hence, the expected signal- yields a true planetary mass M = 0.30+0.14 M . p 0.05 Jup to-noise ratio of the planetary signal is [0.5, 40] (peak-to-peak). In this paper, we use CoRoT observations− to search for light from the planetary companion. We present successively the Therefore, in terms of photon noise, the precision of CoRoT ffi expected performance (Sect. 2), then, the data analysis in the should be su cient, except for a combination of the lower Fourier and the time domain to ensure an ambiguous detection bounds of both inclination and albedo. Note that we have not ( of the planetary light (Sect. 3). In Sect. 4, we identify a spuri- investigated the case where the reflection is anisotropic e.g. ous systematic signal in the CoRoT light curve and propose an Guillot & Vannier 2003). A hypothetical non-strictly sinusoidal approximate correction. We conclude that we achieve a possible planetary signal would include harmonics of the orbital fre- detection of the planetary contribution to the light curve, which quency. Since their values are a priori unknown, and given that would need further and longer observations to strengthen the re- the stellar signal itself displays components of relatively high sult (Sect. 6). and variable amplitudes at frequencies higher than the orbital one, we estimate that the planetary effect will only be detectable in terms of its main orbital frequency (Eq. 1). 2. Expected signal

We neglected the eccentricity of the orbit and assumed that the 3. Ambiguous result from data analyses starlight was reflected isotropically over the “dayside” of the planet (Lambert sphere). The peak-to-peakamplitude of the pho- 3.1. No clear planetary signature in the Fourier spectrum tometric fluctuations caused by the planet is The CoRoT light curve (Fig. ??) does not exhibit signs of high 2 ∆Fp(t) 2A R activity. Only relative variations smaller than 400 ppm are ob- = sin i, (1) F (t) 3 a served on typical timescales of 10 days. On the one hand, the ⋆ lack of activity aids the identification of the planetary signal em- where Fp and F⋆(t) are the planetary and stellar fluxes, A and bedded in the lightcurve.On the other hand, the low activity cou- R the planetary Bond albedo and radius, a the orbital distance, pled with the long stellar rotation period ensures that it is impos- and i the orbitalinclination (e.g. Charbonneau et al. 1999). There sible for the spot analysis to be performed in a similar way to are therefore three degenerate parameters: the albedo, the orbital Mosser et al. (2009). Therefore the stellar rotation estimate can- inclination, and the planetary radius. not be refined, complicating the disentangling of the planetary The planetary radius may be estimated from evolution calcu- from the stellar signal. lations that are parameterized to reproduce the measured radii of We first search for signatures of the planetary changing transiting exoplanets (Guillot 2008). The result depends mostly phases in the Fourier spectrum. From Butler et al. (2006), the on the unknown core mass: for an assumed planetary mass of planetary orbital period is 3.023573 0.000065 day, which cor- ± 0.3MJup and a dissipation of 1% of the incoming stellar light responds to the frequency 0.330735 0.000007 cycles per day at the planet’s center, we derive planetary radii ranging from (c/d). The frequency resolution of a± 34-day run is 0.029 c/d. 0.8 RJup for a 60M dense core to 1.6 RJup for a planet with no Hence, the planetary signal should appear as a peak in the fre- core. ⊕ quency range [0.32, 0.35] c/d. In Fig. 1, we present the power By assuming that the orbital plane is tilted by i = 45◦, density spectrum calculated using the discrete Fourier trans- we found the radius Rp = 1.1 RJup and given the large as- form (DFT). The background, whose amplitude is the result of sumed uncertainty in the planetary albedo, A = 0.05 to 0.5, the stellar activity, was reproduced by fitting a sum of 2 semi- Gaulme et al.: reflected light from HD 46375b 3

30 60

20 40 10

Φ (ppm) 0 / 20

∆Φ −10

−20 0 0 0.5 1 1.5 2 2.5 3 Φ (ppm)

Phase (days) / −20 Fig. 2. Folded time series over 3.023573 days. The 2 red sine ∆Φ curves represent the minimum and maximum limits of the expected signal according to the ephemeris of Butler et al. 2006. −40

30 −60 20

10 0 0.5 1 1.5 2 2.5 3 Phase (days) 0 Φ (ppm) / −10 Fig. 4. Top: folding of HD 46375 time series over the plane- ∆Φ tary period. Middle: folding of HD 46179 over the same period. −20 HD 46375 HD 46179 Bottom: folding of the time series made of the difference be- −30 tween HD 46375 and HD 46179 data. The red sine curves in- 0 0.5 1 1.5 2 2.5 3 Phase (days) dicate the data fitting by excluding the regions indicated in Fig. 3. Fig. 3. Folding of the HD 46375 and HD 46179 time series over 3 days. The grey zones correspond to region where the correla- tion between both light curves is maximum. 4. Evidence of a signal compatible with the planet 4.1. A 24-h periodic systematic signal Lorentzian functions to the whole power spectrum (Gaulme et The peculiarity and difficulty of HD 46375bis that its orbital pe- al. 2010a). riod is close to an integernumberof days, which makes its detec- No excess power appears in the expected frequency range. tion very sensitive to telluric systematic signals. Contaminating Moreover, we note the peak at 0.38 c/d, whose origin is not signals with 1-day periods were shown to be present in the explained by typical systematic signals of the CoRoT, such as CoRoT exoplanetary lightcurves (Mazeh et al. 2009). the orbital period (13.97 c/d), the Earth rotation period (visible In our case, among the 9 stars followed up by CoRoT in around 1 c/d), or a combination of them. Therefore, the Fourier the same observation run in the seismic field, only HD 46179 analysis is inconclusive and a direct analysis of the time se- (mHip = 6.68) exhibits less activity than HD 46375, with varia- ries has to be performed, by taking into account the information tions below 165 ppm after removing a linear trend. We processed about the planetary orbital phase. the light curve in the same way as HD 46375, and then folded both time series exactly over 3 days (Fig 3). The 2 curves appear to be strongly correlated with a clear 24-h period. Therefore, a 3.2. A sine trend in the folded time series common systematic signal is present in both lightcurves. This The alternative way to search for the planetary signal, for which feature is probably related to the CoRoT orbit, and may be both phase and period are known, consists of folding the time caused by e.g. flux variations generated by the entrance/exit in series over the planetary period. To increase the signal-to-noise the night side of the Earth (Auvergne et al. 2009). We note that ratio, we first rebinnedthe data over 2 CoRoT orbital periods, i.e. only HD 46375 exhibits a sine trend when folded about the plan- 6184 s. Then, the low frequency trend, consisting of the moving etary period, which indicated that the sine trend is characteristic average weighted by a triangular window of 7-day FWHM, is of HD 46375 (Fig. 4). subtracted from the time series. We note that we checked this ff step does not a ect the planetary period. The rebinned and flat- 4.2. Cleaning of the time series tened time series is folded over the planetary period. If the noise is dominated by photon noise, the standard deviation of the time We first tried to clean the lightcurves using the SYS-REM algo- series is 176 ppm (Fig. ??) for 92140 data measurements gives rithm (Guterman & Mazeh, private comm.). However, this did for the folded time series (43 bins) has a 3.9 ppm noise level. not lead to a reduction in the standard deviation at low frequen- The analysis of the folded time series over the planetary orbit cies. We therefore adopted a simple cleaning method, by con- detects two main trends (Fig. 2). Firstly, a sine trend with a peak- sidering the HD 46179 data as reference. Firstly, we flattened to-peak amplitude of about 20 ppm appears to be in phase with the HD 46179 lightcurve by removing a linear trend, and then the expected signal. Secondly, a sharp excess power is visible in subtracted it from the HD 46375 time series. The rebinned and the range t = [2.2, 2.5]days, which has an amplitude of about 35 folded time series exhibits a clearer sine trend, while the relative ppm. We must explain the origin of the tall sharp peak, to before amplitude with respect to the sine amplitude is reduced by about we can be certain that the sine trend is not an artefact. 2 (Fig. 4). 4 Gaulme et al.: reflected light from HD 46375b

400 by the CoRoT transmission function (Auvergne et al. 2009). We HD 46375 found that the minimum temperature is at least [1880, 2030] K ) HD 46375 − HD 46179 1

− 300 Background fitting for ∆Fp/F⋆ = [13.0, 26.8] ppm, by assuming that the night side

Hz has a negligible emission and the planet is seen almost edge- µ

2 200 on. This is unlikely, because the ratio of dayside optical brightness temperature to equilibrium temperature would be at least 100 between 1.6 and1.8 for HD 46375b compared to measured val- P (ppm ues of between 1.2and 1.45 for OGLE-TR-56b, CoRoT-1b, and 0 CoRoT-2b. 0.2 0.4 0.6 0.8 1 1.2 ν (cycles per day)

Fig. 5. Power density spectra of the time series corresponding 5. Discussion and conclusion to the difference between HD 46375 and HD 46179 lightcurves, Our analysis of observation data presented here has been the compared to HD 46375. first attempt to directly detect the stellar light reflected by a non-transiting planet in the visible domain. In parallel to the present work, we have performed a seismic analysis on the same Moreover, the Fourier analysis confirms that while the spec- data set. Coupled with ground-based spectropolarimetric obser- trum amplitude decreases at both 1 c/d and 0.5 c/d, it increases vations, obtained with the NARVAL spectropolarimeter at the significantly in the frequency range which including both the ter- Pic du Midi Observatory, we have been able to strongly refine restrial harmonics (1/3c/d) and HD 46375b orbital period (Fig. the stellar modelling (Gaulme et al. 2010a). 5). However, since the light due to Earth-shine is unlikely to fol- We used two approaches to detect the planetary contribution low a simple path to reach the detector, we cannot reject the hy- to the CoRoT light curve.On the one hand, the orbital signal was pothesis that the increase in power at 0.33 c/d could be an arte- only barely detected by our Fourier analysis. On the other hand, fact, introduced by the comparison of two light curves obtained the time series analysis, consisting of rebinning and folding the on different parts of the detector. light curve over the planetary orbital period, has allowed us to detect a signal compatible with the planetary orbital signature. 4.3. A signal compatible with HD 46375b The folded light curve appears to be composed of two mains signals: a sine trend, in phase with ephemeris, and a high excess The variability in the 1-day period spurious signal in both time of intensity, which has no a priori explanation. and space preventsus from properlymodeling it and cleaning the We compared our data with the time series of HD 46179, data. From Fig. 3, the intensity of this spurious signal reaches a which is the only star among those of the same seismic CoRoT maximum in 0.3-day long intervals close to phases [0.3, 1.3, 2.3] run to exhibit a low activity, which is actually lower than that days. We therefore fitted the folded time series by excluding of HD 46375. The 3-day folding of both time series enabled us these 3 ranges, to minimise the influence of the systematics on to detect a common 1-day period systematic signal, with 35 the sine fitting. The fitting function is a simple sine, whose phase ppm maximum amplitude (Fig. 3). We reduced by a factor≈ 2 the φ, amplitude ∆Fp/F⋆, and offset are free. The harmonics of the relative amplitude of the spurious peak with respect to the sine planetary orbit are left aside, because of the low signal-to-noise trend, by subtracting the filtered time series of HD 46179 from level. the HD 46375 time series. We first fitted the raw, then the “roughly cleaned” light We illustrated that the data is affected by the presence of a curves. The least squares fitting (reduced χ2 = 1.75) measured systematic signal presumedly of telluric origin, which is strongly a relative peak-to-valley amplitude ∆Fp/F⋆ = 13.0 5.0 ppm variable in both time and space. It appears necessary to study ± and a phase shift of 19 9◦ with respect to the ephemeris. A the photometric fluctuations at low frequency in the whole 3- ± phase shift of 19◦ corresponds to 1.4-σ with the Butler et al. year CoRoT set data especially in the seismic field, to refine the (2006) error bar on the elongation date, which is reasonable modeling of spurious signals. This objective goes far beyond the since their value lies on 50 measurements, all performed before aims of the present work. However, the only way of determin- 2006. The fitting of the “cleaned” data allows us to improve ing whether that the sine trend is of planetary origin consists of the fitting (χ2 = 0.87) and to re-estimate the sine amplitude observing the HD 46375 system for a duration long enough to ∆Fp/F⋆ = 26.8 3.0 ppm, with a phase shift φ = 29 6◦, i.e. disentangle the planetary orbital period from 3 days, i.e. a dura- ± ± corresponding to a 2-σ error relative to the predicted ephemeris. tion longer than 100 days, as achievable with a CoRoT long Moreover, the amplitude of the peak corresponding to the plan- run. ∼ etary period is significantly larger in the density spectrum of the cleaned data (Fig. 5). After fitting the folded time series by excluding three References regions, we identified a signal compatible with the planet. Alonso, R., Alapini, A., Aigrain, S., et al. 2009a, A&A, 506, 353 Assuming that i = 45◦, R = 1.1 RJup, and ∆Fp/F⋆ = Alonso, R., Guillot, T., Mazeh, T., et al. 2009b, A&A, 501, L23 [13.0, 26.8] ppm, our inferred albedo is between 0.16 and Auvergne, M., Bodin, P., Boisnard, L., et al. 2009, A&A, 506, 411 0.33, which is higher than expected on the basis of theoretical Baglin, A., Auvergne, M., Barge, P., et al. 2002, in ESA Special Publication, Vol. 485, Stellar Structure and Habitable Planet Finding, ed. B. Battrick, F. Favata, models, but within the range of actually derived albedos (see I. W. Roxburgh, & D. Galadi, 17–24 Cowan & Agol 2010). Slightly larger radii and inclinations, of Borucki, W. J., Koch, D., Jenkins, J., et al. 2009, Science, 325, 709 course, yield lower albedos consistent with theoretical expec- Butler, R. P., Wright, J. T., Marcy, G. W., et al. 2006, ApJ, 646, 505 tations (see Sudarsky et al. 2003). Alternatively, the observed Charbonneau, D., Noyes, R. W., Korzennik, S. G., et al. 1999, ApJ, 522, L145 Cowan, N. B. & Agol, E. 2010, ArXiv e-prints flux may be produced by direct thermal emission. The bright- Guillot, T. 2008, Physica Scripta Volume T, 130, 014023 ness temperatures can be calculated from the ratio of the black Guillot, T. & Vannier, M. 2003, in EAS Publications Series, Vol. 8, EAS body emissions of the planet with respect to the star, weighted Publications Series, ed. C. Aime & R. Soummer, 25–35 Gaulme et al.: reflected light from HD 46375b 5

Harrington, J., Hansen, B. M., Luszcz, S. H., et al. 2006, Science, 314, 623 Hebrard, G., Desert, J., Diaz, R. F., et al. 2010, ArXiv e-prints Henry, G. W. 2000, ApJ, 536, L47 Marcy, G. W., Butler, R. P., & Vogt, S. S. 2000, ApJ, 536, L43 Mazeh, T., Guterman, P., Aigrain, S., et al. 2009, A&A, 506, 431 Mosser, B., Baudin, F., Lanza, A. F., et al. 2009, A&A, 506, 245 Seager, S., Whitney, B. A., & Sasselov, D. D. 2000, ApJ, 540, 504 Showman, A. P. & Guillot, T. 2002, A&A, 385, 166 Sing, D. K. & L´opez-Morales, M. 2009, A&A, 493, L31 Snellen, I. A. G., de Mooij, E. J. W., & Albrecht, S. 2009, Nature, 459, 543 Sudarsky, D., Burrows, A., & Hubeny, I. 2003, ApJ, 588, 1121 800

600

400

200

−200

−400

−600 Photometric variations (ppm) −800 Time series Smoothed time series −1000 0 5 10 15 20 25 30 Time (days)