ARIEL: Science, Mission & Community 2020 Conférence - ESTEC - January 2020
SYNERGIES BETWEEN RADIAL VELOCITIES AND THE ARIEL MISSION
ALEXANDRE SANTERNE
AIX-MARSEILLE UNIVERSITY / LABORATOIRE D’ASTROPHYSIQUE DE MARSEILLE
@A_SANTERNE ARIEL: Science, Mission & Community 2020 Conférence - ESTEC - January 2020
SYNERGIES BETWEEN RADIAL needs RVs ? VELOCITIESWhy AND ARIEL THE ARIEL MISSION
ALEXANDRE SANTERNE
AIX-MARSEILLE UNIVERSITY / LABORATOIRE D’ASTROPHYSIQUE DE MARSEILLE
@A_SANTERNE What ARIEL wants as input:
• The mass of exoplanets
• Precise ephemerides for both the primary and secondary transits
• To know the host star activity
kT Scale Height H = µmg
kT Scale Height H = Planet mass µmg
kT Scale Height H = Planet mass µmg
m sin i K p 1 2 / P 3 M 3 p1 e2 ?
Mayor & Queloz (1995) Required Precision of Exoplanet Masses 5
Mass of (giant) exoplanets might be retrieved from transmission -10 -5 0 5 10 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 spectroscopic data Temperature (% from truth ) [M/H] - [M/H]truth (log Solar) De Wit & Seager (2013)
-8 -6 -4 -2 0 2 4 6 8 -40 -30 -20 -10 0 10 20 30 40 xRp (% from truth) Mass (% from truth)
w/ Mass WASP 17 b HAT-P-1 b WASP-12 b HAT-P-26 b
No Mass Hot Jupiters Hot Neptune
Figure 2. Retrieved posterior distributions for the three hot Jupiters (WASP-17b, HAT-P-1b, and HD 189733 b) and the hot Neptune (HAT-P-26 b). The four retrieved parameters shown here (temperature,Batalha et metallicity, al. (2017, reference 2019) radius, and mass) are a subset of the full state vector. All posteriors are shown relative to the true value. The shaded posteriors are for the retrieval cases where mass was fixed to the known value and the lines are when mass was a free parameter. Main Point: For the hot Jupiters, we find that the constraint on the atmospheric parameters is not largely dependent on whether or not mass is included as a free parameter. dius and atmospheric parameters. We do find a degen- licity (4 Solar) compared with the hot Jupiter-cases ⇥ eracy between reference radius and mass that leads to (1 Solar) and warm-Neptune-cases (>400 Solar). ⇥ ⇥ slightly overestimated mass constraints for two of the Therefore, it o↵ers an opportunity to determine the hot Jupiters (+6.6% for WASP-17b and +5.8% HAT- e↵ect of small changes in metallicity on the widths of T-1b), and underestimation for the third hot Jupiter the retrieved posterior distributions of mass, reference ( -12.5% for WASP-12b). This degeneracy is not sur- radius and the atmospheric parameters. While the prising given g = GM/R(z)2 and any o↵set in reference precision of the retrieved metallicity and temperature radius leads to a slight o↵set in retrieved mass. distributions is not a↵ected by an unknown mass, the Out of the three hot Jupiters, WASP-12b is the case accuracy of both are slightly degraded. Additionally, with the highest degree of cloud coverage. This leads the accuracy and the precision of the retrieved reference to spectral features that are highly muted, relative to radius are both highly impacted by the unknown mass. WASP-17b and HAT-P-1b. Cloud-coverage is a well- The +4% overestimation of the reference radius leads to documented e↵ect that can sometimes impede precise a +20% overestimation of the mass. This degeneracy be- retrievals of atmospheric parameters (B´etr´emieux 2016; tween mass and radius was seen as a minor e↵ect in the B´etr´emieux & Swain 2017; Line & Parmentier 2016; hot Jupiter cases where metallicity was fixed at 1 Solar. ⇥ MacDonald & Madhusudhan 2017). Therefore, WASP- This foreshadows higher mean molecular weight as a po- 12b exhibits wider retrieved posteriors for temperature, tential inhibitor to constraining atmospheric properties, metallicity, reference radius, and mass, as shown in Fig- including mass, when mass is unknown. ure 2, relative to the other two hot Jupiters. Figure 2 also shows the case for hot Neptune HAT- P-26b. HAT-P-26b has a slightly enhanced metal- 3.2. GJ 436b & GJ 1214b Required Precision of Exoplanet Masses 5
Mass of (giant) exoplanets might be retrieved from transmission -10 -5 0 5 10 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 spectroscopic data Temperature (% from truth ) [M/H] - [M/H]truth (log Solar) De Wit & Seager (2013)
THIS MIGHT INTRODUCE BIAS IN THE STATISTICAL RESULTS -8 -6 -4 -2 0 2 4 6 8 -40 -30 -20 -10 0 10 20 30 40 xRp (% from truth) Mass (% from truth)
w/ Mass WASP 17 b HAT-P-1 b WASP-12 b HAT-P-26 b
No Mass Hot Jupiters Hot Neptune
Figure 2. Retrieved posterior distributions for the three hot Jupiters (WASP-17b, HAT-P-1b, and HD 189733 b) and the hot Neptune (HAT-P-26 b). The four retrieved parameters shown here (temperature,Batalha et metallicity, al. (2017, reference 2019) radius, and mass) are a subset of the full state vector. All posteriors are shown relative to the true value. The shaded posteriors are for the retrieval cases where mass was fixed to the known value and the lines are when mass was a free parameter. Main Point: For the hot Jupiters, we find that the constraint on the atmospheric parameters is not largely dependent on whether or not mass is included as a free parameter. dius and atmospheric parameters. We do find a degen- licity (4 Solar) compared with the hot Jupiter-cases ⇥ eracy between reference radius and mass that leads to (1 Solar) and warm-Neptune-cases (>400 Solar). ⇥ ⇥ slightly overestimated mass constraints for two of the Therefore, it o↵ers an opportunity to determine the hot Jupiters (+6.6% for WASP-17b and +5.8% HAT- e↵ect of small changes in metallicity on the widths of T-1b), and underestimation for the third hot Jupiter the retrieved posterior distributions of mass, reference ( -12.5% for WASP-12b). This degeneracy is not sur- radius and the atmospheric parameters. While the prising given g = GM/R(z)2 and any o↵set in reference precision of the retrieved metallicity and temperature radius leads to a slight o↵set in retrieved mass. distributions is not a↵ected by an unknown mass, the Out of the three hot Jupiters, WASP-12b is the case accuracy of both are slightly degraded. Additionally, with the highest degree of cloud coverage. This leads the accuracy and the precision of the retrieved reference to spectral features that are highly muted, relative to radius are both highly impacted by the unknown mass. WASP-17b and HAT-P-1b. Cloud-coverage is a well- The +4% overestimation of the reference radius leads to documented e↵ect that can sometimes impede precise a +20% overestimation of the mass. This degeneracy be- retrievals of atmospheric parameters (B´etr´emieux 2016; tween mass and radius was seen as a minor e↵ect in the B´etr´emieux & Swain 2017; Line & Parmentier 2016; hot Jupiter cases where metallicity was fixed at 1 Solar. ⇥ MacDonald & Madhusudhan 2017). Therefore, WASP- This foreshadows higher mean molecular weight as a po- 12b exhibits wider retrieved posteriors for temperature, tential inhibitor to constraining atmospheric properties, metallicity, reference radius, and mass, as shown in Fig- including mass, when mass is unknown. ure 2, relative to the other two hot Jupiters. Figure 2 also shows the case for hot Neptune HAT- P-26b. HAT-P-26b has a slightly enhanced metal- 3.2. GJ 436b & GJ 1214b Required Precision of Exoplanet Masses 5
Mass of (giant) exoplanets might be retrieved from transmission -10 -5 0 5 10 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 spectroscopic data Temperature (% from truth ) [M/H] - [M/H]truth (log Solar) De Wit & Seager (2013)
THIS MIGHT INTRODUCE BIAS IN THE STATISTICAL RESULTS -8 -6 -4 -2 0 2 4 6 8 -40 -30 -20 -10 0 10 20 30 40 xRp (% from truth) Mass (% from truth)
w/ Mass WASP 17 b HAT-P-1 b WASP-12 b HAT-P-26 b
Precise masses from RVs are still relevant No Mass Hot Jupiters Hot Neptune in the context of ARIEL ! Figure 2. Retrieved posterior distributions for the three hot Jupiters (WASP-17b, HAT-P-1b, and HD 189733 b) and the hot Neptune (HAT-P-26 b). The four retrieved parameters shown here (temperature,Batalha et metallicity, al. (2017, reference 2019) radius, and mass) are a subset of the full state vector. All posteriors are shown relative to the true value. The shaded posteriors are for the retrieval cases where mass was fixed to the known value and the lines are when mass was a free parameter. Main Point: For the hot Jupiters, we find that the constraint on the atmospheric parameters is not largely dependent on whether or not mass is included as a free parameter. dius and atmospheric parameters. We do find a degen- licity (4 Solar) compared with the hot Jupiter-cases ⇥ eracy between reference radius and mass that leads to (1 Solar) and warm-Neptune-cases (>400 Solar). ⇥ ⇥ slightly overestimated mass constraints for two of the Therefore, it o↵ers an opportunity to determine the hot Jupiters (+6.6% for WASP-17b and +5.8% HAT- e↵ect of small changes in metallicity on the widths of T-1b), and underestimation for the third hot Jupiter the retrieved posterior distributions of mass, reference ( -12.5% for WASP-12b). This degeneracy is not sur- radius and the atmospheric parameters. While the prising given g = GM/R(z)2 and any o↵set in reference precision of the retrieved metallicity and temperature radius leads to a slight o↵set in retrieved mass. distributions is not a↵ected by an unknown mass, the Out of the three hot Jupiters, WASP-12b is the case accuracy of both are slightly degraded. Additionally, with the highest degree of cloud coverage. This leads the accuracy and the precision of the retrieved reference to spectral features that are highly muted, relative to radius are both highly impacted by the unknown mass. WASP-17b and HAT-P-1b. Cloud-coverage is a well- The +4% overestimation of the reference radius leads to documented e↵ect that can sometimes impede precise a +20% overestimation of the mass. This degeneracy be- retrievals of atmospheric parameters (B´etr´emieux 2016; tween mass and radius was seen as a minor e↵ect in the B´etr´emieux & Swain 2017; Line & Parmentier 2016; hot Jupiter cases where metallicity was fixed at 1 Solar. ⇥ MacDonald & Madhusudhan 2017). Therefore, WASP- This foreshadows higher mean molecular weight as a po- 12b exhibits wider retrieved posteriors for temperature, tential inhibitor to constraining atmospheric properties, metallicity, reference radius, and mass, as shown in Fig- including mass, when mass is unknown. ure 2, relative to the other two hot Jupiters. Figure 2 also shows the case for hot Neptune HAT- P-26b. HAT-P-26b has a slightly enhanced metal- 3.2. GJ 436b & GJ 1214b ARIEL wants precisely the ephemerides of transits ARIEL wants precisely the ephemerides of transits
• By the time ARIEL will be launched, many transit ephemerides will be lost ARIEL wants precisely the ephemerides of transits
• By the time ARIEL will be launched, many transit ephemerides will be lost
• Amateur astronomers might help to maintain the ephemerides
Exoplanet Transit Database ARIEL wants precisely the ephemerides of transits
• By the time ARIEL will be launched, many transit ephemerides will be lost
• Amateur astronomers might help to maintain the ephemerides
• RVs might also contribute for the long-period planets (P>10d) or the shallowest transits Exoplanet Transit Database Improving ephemerides: the CHEOPS example
• CHEOPS transit-search program targeting RV-discovered exoplanets also needs precise ephemerides
• Few SOPHIE re-observations substantially improved predicted transit ephemerides Improving ephemerides: the CHEOPS example
• CHEOPS transit-search program targeting RV-discovered exoplanets also
5.2.needs LA RECHERCHE precise ephemerides EN TRANSIT DE PLANETES` CARACTERIS´ EES´ EN VITESSES RADIALES 155 •TableFew5.4 SOPHIE – Eph´ em´ erides´ re-observations des planetes` observ substantiallyees´ par SOPHIE. Lesimproved erreurs sur predicted la periodes´ de 0.0000transit sont enephemerides realit´ e´ inferieures´ a` 0.0001. Ceci est duˆ au format de sortie des parametres` de DACE qui ne stocke pas les nombres plus petits.
PlanPlanetete` NombreNumber d’observationsof observations LastDerni observationere` obs [BJD] T0 [BJD] P [jour]days HD 46974 c 63 (HIRES) 53753.85 58119.14 3.1 4.9493 0.0002 + 21 (SOPHIE) 57468.41 58123.18 ±0.05 4.9474 ±0.00000 ± ± HD 3651 b (k1) 68 (HIRES) 52608.73 58090.41 1.85 62.2706 0.018 + 15 (SOPHIE) 57417.30 58089.13± 0.7 62.2548 ±0.0072 ± ± Gl 176 b 98 (HIRES et HARPS) 54545.77 58122.54 0.34 8.7758 0.0006 + 25 (SOPHIE) 57468.35 58122.32 ± 0.24 8.7754 ± 0.0004 ± ±
Table 5.5 – Analyse des donnBastienees´ combin Courcolees´ 2016 HIRES (PhD thesis)+ SOPHIE pour HD 49674 b
Parametres` HD 49674 b P [jours] 4.9474 (< 0.0001) ± T p [BJD] 58121.86 0.24 e 0.09 0±.03 ω [deg] 15.16± 17.9 K[ms 1]− 13.94 ±0.36 − ± Msini [M ] 36.99 1.55 ⊕ 1 ± σ (O - C) HIRES [ m s− ] 5.84 1 σ (O - C) SOPHIE [ m s− ] 3.71
et sur l’époque de transit a été grandement améliorée. L’erreur à 1σ sur T0 est inférieure au jour, l’objectif principal du programme a donc bien été atteint. Les cas individuels sont détaillés dans les paragraphes suivants.
HD 49674 b HD 49674 b est un Neptune massif en orbite courte (4.94 jours) autour d’une étoile G0. La Table 5.5 regroupe les paramètres physiques et orbitaux de HD 49674 b et la Figure 5.14 montre les vitesses radiales repliées en phases pour cette planète. On peut voir que les données SOPHIE ont des résidus encore une fois plus faible que les mesures HIRES. D’autre part, les résultats de cette analyse jointe est cohérente avec les données publiées par Butler et al. (2006), avec une précision accrue sur l’ensemble des paramètres. La masse obtenue est légèrement supérieure à 36.99 1.55 M contre 32.3 2.6M pour Butler et al. (2006). ± ⊕ ± ⊕ La Figure 5.15 montre la distribution postérieure de l’époque de transit T0 avec et sans les données SOPHIE. Sans les données SOPHIE, ce paramètre n’est pas bien contrainte avec un tel écart temporel et le modèle saute régulièrement d’une solution orbitale à l’autre à T0 P. Il s’agit d’un problème du MCMC de DACE, en cours de résolution. Cela n’est pas le cas avec les± données SOPHIE, qui permettent d’obtenir le temps de transit avec une très bonne précision. Comme on peut le voir à la Figure 5.14, cela peut s’expliquer par une très bonne couverture de la phase de cette planète avec SOPHIE. Il s’agit d’un exemple typique de l’intérêt de ces observations préparatoires.
HD 3651 b et c HD 3651 est un système à deux planètes, avec un Saturne à 62 jours (Wittenmyer et al., Improving ephemerides: the CHEOPS example
• CHEOPS transit-search program targeting RV-discovered exoplanets also
5.2.needs LA RECHERCHE precise ephemerides EN TRANSIT DE PLANETES` CARACTERIS´ EES´ EN VITESSES RADIALES 155 •TableFew5.4 SOPHIE – Eph´ em´ erides´ re-observations des planetes` observ substantiallyees´ par SOPHIE. Lesimproved erreurs sur predicted la periodes´ de 0.0000transit sont enephemerides realit´ e´ inferieures´ a` 0.0001. Ceci est duˆ au format de sortie des parametres` de DACE qui ne stocke pas les nombres plus petits.
PlanPlanetete` NombreNumber d’observationsof observations LastDerni observationere` obs [BJD] T0 [BJD] P [jour]days HD 46974 c 63 (HIRES) 53753.85 58119.14 3.1 4.9493 0.0002 + 21 (SOPHIE) 57468.41 58123.18 ±0.05 4.9474 ±0.00000 ± ± HD 3651 b (k1) 68 (HIRES) 52608.73 58090.41 1.85 62.2706 0.018 + 15 (SOPHIE) 57417.30 58089.13± 0.7 62.2548 ±0.0072 ± ± Gl 176 b 98 (HIRES et HARPS) 54545.77 58122.54 0.34 8.7758 0.0006 + 25 (SOPHIE) 57468.35 58122.32 ± 0.24 8.7754 ± 0.0004 ± ±
Table 5.5 – Analyse des donnBastienees´ combin Courcolees´ 2016 HIRES (PhD thesis)+ SOPHIE pour HD 49674 b
Parametres` HD 49674 b P [jours] 4.9474 (< 0.0001) ± T p [BJD] 58121.86 0.24 e 0.09 0±.03 ω [deg] 15.16± 17.9 K[ms 1]− 13.94 ±0.36 − ± Msini [M ] 36.99 1.55 ⊕ 1 ± σ (O - C) HIRES [ m s− ] 5.84 1 σ (O - C) SOPHIE [ m s− ] 3.71
et sur l’époque de transit a été grandement améliorée. L’erreur à 1σ sur T0 est inférieure au jour, l’objectif principal du programme a donc bien été atteint. Les cas individuels sont détaillés dans les paragraphes suivants.
HD 49674 b HD 49674 b est un Neptune massif en orbite courte (4.94 jours) autour d’une étoile G0. La Table 5.5 regroupe les paramètres physiques et orbitaux de HD 49674 b et la Figure 5.14 montre les vitesses radiales repliées en phases pour cette planète. On peut voir que les données SOPHIE ont des résidus encore une fois plus faible que les mesures HIRES. D’autre part, les résultats de cette analyse jointe est cohérente avec les données publiées par Butler et al. (2006), avec une précision accrue sur l’ensemble des paramètres. La masse obtenue est légèrement supérieure à 36.99 1.55 M contre 32.3 2.6M pour Butler et al. (2006). ± ⊕ ± ⊕ La Figure 5.15 montre la distribution postérieure de l’époque de transit T0 avec et sans les données SOPHIE. Sans les données SOPHIE, ce paramètre n’est pas bien contrainte avec un tel écart temporel et le modèle saute régulièrement d’une solution orbitale à l’autre à T0 P. Il s’agit d’un problème du MCMC de DACE, en cours de résolution. Cela n’est pas le cas avec les± données SOPHIE, qui permettent d’obtenir le temps de transit avec une très bonne précision. Comme on peut le voir à la Figure 5.14, cela peut s’expliquer par une très bonne couverture de la phase de cette planète avec SOPHIE. Il s’agit d’un exemple typique de l’intérêt de ces observations préparatoires.
HD 3651 b et c HD 3651 est un système à deux planètes, avec un Saturne à 62 jours (Wittenmyer et al., Improving ephemerides: the CHEOPS example
• CHEOPS transit-search program targeting RV-discovered exoplanets also
5.2.needs LA RECHERCHE precise ephemerides EN TRANSIT DE PLANETES` CARACTERIS´ EES´ EN VITESSES RADIALES 155 •TableFew5.4 SOPHIE – Eph´ em´ erides´ re-observations des planetes` observ substantiallyees´ par SOPHIE. Lesimproved erreurs sur predicted la periodes´ de 0.0000transit sont enephemerides realit´ e´ inferieures´ a` 0.0001. Ceci est duˆ au format de sortie des parametres` de DACE qui ne stocke pas les nombres plus petits.
PlanPlanetete` NombreNumber d’observationsof observations LastDerni observationere` obs [BJD] T0 [BJD] P [jour]days HD 46974 c 63 (HIRES) 53753.85 58119.14 3.1 4.9493 0.0002 + 21 (SOPHIE) 57468.41 58123.18 ±0.05 4.9474 ±0.00000 ± ± HD 3651 b (k1) 68 (HIRES) 52608.73 58090.41 1.85 62.2706 0.018 + 15 (SOPHIE) 57417.30 58089.13± 0.7 62.2548 ±0.0072 ± ± Gl 176 b 98 (HIRES et HARPS) 54545.77 58122.54 0.34 8.7758 0.0006 + 25 (SOPHIE) 57468.35 58122.32 ± 0.24 8.7754 ± 0.0004 ± ±
Table 5.5 – Analyse des donnBastienees´ combin Courcolees´ 2016 HIRES (PhD thesis)+ SOPHIE pour HD 49674 b
Parametres` HD 49674 b P [jours] 4.9474 (< 0.0001) ± T p [BJD] 58121.86 0.24 e 0.09 0±.03 ω [deg] 15.16± 17.9 K[ms 1]− 13.94 ±0.36 − ± Msini [M ] 36.99 1.55 ⊕ 1 ± σ (O - C) HIRES [ m s− ] 5.84 1 σ (O - C) SOPHIE [ m s− ] 3.71
et sur l’époque de transit a été grandement améliorée. L’erreur à 1σ sur T0 est inférieure au jour, l’objectif principal du programme a donc bien été atteint. Les cas individuels sont détaillés dans les paragraphes suivants.
HD 49674 b HD 49674 b est un Neptune massif en orbite courte (4.94 jours) autour d’une étoile G0. La Table 5.5 regroupe les paramètres physiques et orbitaux de HD 49674 b et la Figure 5.14 montre les vitesses radiales repliées en phases pour cette planète. On peut voir que les données SOPHIE ont des résidus encore une fois plus faible que les mesures HIRES. D’autre part, les résultats de cette analyse jointe est cohérente avec les données publiées par Butler et al. (2006), avec une précision accrue sur l’ensemble des paramètres. La masse obtenue est légèrement supérieure à 36.99 1.55 M contre 32.3 2.6M pour Butler et al. (2006). ± ⊕ ± ⊕ La Figure 5.15 montre la distribution postérieure de l’époque de transit T0 avec et sans les données SOPHIE. Sans les données SOPHIE, ce paramètre n’est pas bien contrainte avec un tel écart temporel et le modèle saute régulièrement d’une solution orbitale à l’autre à T0 P. Il s’agit d’un problème du MCMC de DACE, en cours de résolution. Cela n’est pas le cas avec les± données SOPHIE, qui permettent d’obtenir le temps de transit avec une très bonne précision. Comme on peut le voir à la Figure 5.14, cela peut s’expliquer par une très bonne couverture de la phase de cette planète avec SOPHIE. Il s’agit d’un exemple typique de l’intérêt de ces observations préparatoires.
HD 3651 b et c HD 3651 est un système à deux planètes, avec un Saturne à 62 jours (Wittenmyer et al., Orbital eccentricities and secondary transits Orbital eccentricities and secondary transits
• Eccentric orbit might prevent a transiting planet to be occulted by the star (as seen from the Earth) Orbital eccentricities and secondary transits
• Eccentric orbit might prevent a transiting planet to be occulted by the star (as seen from the Earth)
• Eccentricity of transiting exoplanets is mostly unknown or poorly determined
160 / 850 giant (R>0.3Rjup) planets
source: NASA exoplanet archive Orbital eccentricities HAT-P-29 b and secondary transits
• Eccentric orbit might prevent a transiting planet to be occulted by the star (as seen from the Earth)
• Eccentricity of transiting exoplanets is mostly unknown or poorly determined
Bonomo et al. (2017)
160 / 850 giant (R>0.3Rjup) planets RVs can improve orbital eccentricities of ARIEL targets
source: NASA exoplanet archive Period determination of TESS mono-transit Period determination of TESS mono-transit
• TESS is finding interesting warm giants with periods > 10 days. They transited only once (aka mono-transit) in the 27-days TESS sectors (unless close to CVZ). Period determination of TESS mono-transit
• TESS is finding interesting warm giants with periods > 10 days. They transited only once (aka mono-transit) in the 27-days TESS sectors (unless close to CVZ).
• On-going SOPHIE programme on TESS giant monotransits. First planets
detected ! 100 80 Half Jupiter-mass planet @ 18 days 60
40
20
ΔRV [m/s] ΔRV 0
-20
-40
-60
8760 8780 8800 8820 8840 8860 Date (BJD - 2,450,000.0) [d] Period determination of TESS mono-transit
• TESS is finding interesting warm giants with periods > 10 days. They transited only once (aka mono-transit) in the 27-days TESS sectors (unless close to CVZ).
• On-going SOPHIE programme on TESS giant monotransits. First planets
detected ! 100 80 Half Jupiter-mass planet @ 18 days 60
Next step: use CHEOPS 40 or ground-based 20
ΔRV [m/s] ΔRV 0 observatories to detect a -20 second transit -40 -60
8760 8780 8800 8820 8840 8860 Date (BJD - 2,450,000.0) [d] TESS extension on monotransit TESS extension on monotransit
• During TESS’ extension, a second transit might be detected, a few years after the first transit. It results in a series of possible ephemerides (Δt and all harmonics) TESS extension on monotransit
• During TESS’ extension, a second transit might be detected, a few years after the first transit. It results in a series of possible ephemerides (Δt and all harmonics)
• Detecting the planetary signal with RVs help to select the correct period among all the solutions. TESS extension on monotransit
• During TESS’ extension, a second transit might be detected, a few years after the first transit. It results in a series of possible ephemerides (Δt and all harmonics)
• Detecting the planetary signal with RVs help to select the correct period among all the solutions.
The case of HIP41378 … The case of HIP41378 (V=8.9 - K=7.7)
Vanderburg et al. (2016)
• Detection of 5 transiting planets by K2 in 2015
• Planets b & c are multi-transit while d, e, & f are mono-transit The case of HIP41378 (II)
• Reobservation by K2 in 2018 (3 years later)
New Transits of HIP 41378 11 • Detection of a second transit for planets d & f: 23 different solutions for each planet (519 combinations) following way. First, we exclude all orbital periods gen- erated by Eq. 3 that fall below the lower limit derived TABLE 2 Possible orbital periods for by Eq. 2. Then, for each remaining orbital period, we HIP 41378 d extract the probability from the interpolated product of the baseline prior and the PDF of dynamically feasible Orbital Period (days) Normalized periods (this function is plotted as the solid line in Fig. Probability 4) at exactly that orbital period. We repeat this for each 1113.4465 0.0034 < 0.1 % possible period, and then normalize the total probability 556.7233 ±0.0017 < 0.1 % 371.1488 ± 0.0011 0.1 % for all discrete periods to be equal to one. The resultant ± periods and their corresponding normalized probabilities 278.3616 0.0009 0.5 % 222.6893 ± 0.0007 1.1 % are presented in Table 2 and Table 3. 185.5744 ± 0.0006 2.4 % 159.0638 ± 0.0005 4.1 % 4.4. Final period constraints for HIP 41378 e 139.1808 ± 0.0004 5.7 % HIP 41378 e transited once during K2 C5, but did not 123.7163 ± 0.0004 6.7 % 111.3447 ± 0.0003 7.1 % transit during C18 (see Fig. 1). As such, we do not have 101.2224 ± 0.0003 7.1 % discrete guesses for its true orbital period; however, we 92.7872 ±0.0003 7.0 % can exclude any orbital period that would have led to a 85.6497 ± 0.0003 6.9 % transit being observable during C18. To construct an ad- 79.5319 ± 0.0002 6.8 % 74.2298 ± 0.0002 6.8 % ditional PDF that represents this scenario, we test each 69.5904 ± 0.0002 6.3 % possible orbital period for HIP 41378 e between 72 days 65.4969 ± 0.0002 5.9 % (the minimum orbital period permitted by Eq. [2]) and 61.8581 ± 0.0002 5.5 % 58.6024 ± 0.0002 5.1 % 1200 days. Then, we allow te,18 to vary between the times ± of the first and last data points of C18. If Eq. 3 is satis- 55.6723 0.0002 4.8 % 53.0213 ± 0.0002 4.5 % fied for some integer i for any value of te,18 on this range, 50.6112 ± 0.0002 4.2 % then we consider this particular period “observable” in 48.4107 ± 0.0001 1.4 % ± C18, and set the probability that it is the true orbital Note.—Possibleorbitalsperiods Beckerperiod et of al. HIP (2018) 41378 e; toBerardo zero. The et result al. (2019) of this pruning and their relative likelihoods, based (normalized so the maximum probability is equal to the on the dynamical analysis described in maximum probability of the PDF constructed from the Section4. Values may not add up to 100% due to rounding. Errors on or- results of the baseline PDF and dynamical analysis) is bital periods were computed with = shown in grey in Figure 5. 2 2 1/2 (tc,5 + tc,18) /n,whenn denotes the The final orbital period for HIP 41378 e cannot be di- number of full cycles between C5 and rectly constrained due to the lack of a transit in the C18 C18, and tc denotes the uncertainty data; the best that can be done without follow-up obser- on center time of transit in each cam- paign. Errors on the orbital period are vations is the probabilistic period estimation presented lower when a larger number of periods in the bottom panel of Figure 5. have elapse since the C5 observation. 5. DISCUSSION to produce continuous precise light curves across mul- 5.1. Strategies for observational follow-up in the tiple observing sites around the globe (Boyajian et al. HIP 41378 system 2018), may be well suited to detect the long duration In this work, we have identified a discrete set of pre- transit of HIP 41378 f. The shallower (800 ppm) transits cise possible orbital periods for the long-period transiting of HIP 41378 d, however, will likely require space-based planets HIP 41378 d and HIP 41378 f, and have assessed resources such as the Spitzer Space Telescope, or poten- the likelihood that each of these possible orbital periods tially the CHEOPS space telescope once it launches in is indeed the true orbital period. While we have signif- 2019, for confirmation. icantly constrained the possible orbital periods of these Because of HIP 41378 d’s shorter orbital period, and two planets (we have ruled out about 25% of the pos- the fact that our ground-based data from HAT and sible periods at confidence <1.5% for planet d and 80% KELT were unable to detect or rule out its shallow tran- of possible periods for planet f), our analysis is so far sits, there are a large number of possible orbital periods, unable to uniquely determine the true orbital periods of many of which have roughly equal probabilities of being these planets. Additional follow-up observations will be the true orbital period. Observing transits at all of these necessary to ultimately identify the true orbital periods possible transit times would be an expensive observing and enable future studies with facilities like JWST. program for a precious resource like Spitzer. However, it To determine the true orbital periods for HIP 41378 d should be possible to significantly increase the e ciency and f, the strategy is fairly straightforward. The addi- of Spitzer follow-up observations for these possible or- tional transits during C18 and our identification of pre- bital periods because of how many of these periods are cise possible orbital periods makes it possible to sched- related to one another by harmonics. For example, a sin- ule targeted transit follow-up observations at these most gle Spitzer non-detection of a transit of HIP 41378 d on likely periods. The 0.5% transit depth of HIP 41378 the observation opportunity on 2019 June 16 (371.149 f makes it possible to detect the transit with ground- days after the C18 transit) would rule out four possible based telescopes, although the long (19 hour) transit du- orbital periods (371.149, 185.574, 123.716, and 61.858 ration will make it impossible to observe the transit from days). Taking advantage of these harmonic relationships a single observatory. The multi-site Las Cumbres Obser- between the possible orbital periods may significantly de- vatory telescopes, which have demonstrated the ability crease the amount of observing time needed to identify The case of HIP41378 (III) The case of HIP41378 (III) • Intensive RVs observations (450+ epochs) by 4 stabilised spectrographs over 4 years (HARPS, HARPS-N, HIRES, PFS) a 10 ] 1 0 [m.s
Radial velocity 10 7400 7600 7800 8000 8200 8400 8600 Time [BJD - 2 450 000] b 10 ] 1 0 [m.s
Radial velocity 10 The case of HIP41378 (III) 8400 8450 8500 8550 8600 • Intensive RVs observations (450+ epochs) by 4Timestab [BJDilise -d 2 450 000] spectrographs ovHIP41378er 4 yea brs (HARPS, HARPS-N, HIRES, PFS) HIP41378 c HIP41378 g c d e 2 2 2
• De] tection of the Doppler signals of HIP41478 b, c, and f (+ non- 1 tra nsi0ting planet g) 0 0 [m.s
Radial velocity 2 2 2 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00
HIP41378 d HIP41378 e HIP41378 f af g h 102 2 2 ] ] 1 1 0 0 0 [m.s [m.s
Radial velocity 2 2 2 Radial velocity 10 0.740000 0.25 0.5076000.75 1.0078000.00 0.25 80000.50 0.75 82001.00 0.00 08400.25 0.50 0.860075 1.00 Orbital phase TimeOrbital [BJD - phase 2 450 000] Orbital phase b 10 HARPS HARPS-N HIRES PFS ] 1
Santerne et al. (2020) 0 [m.s
Radial velocity 10 8400 8450 8500 8550 8600 Time [BJD - 2 450 000] HIP41378 b HIP41378 c HIP41378 g c d e 2 2 2 ] 1 0 0 0 [m.s
Radial velocity 2 2 2 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00
HIP41378 d HIP41378 e HIP41378 f f g h 2 2 2 ] 1 0 0 0 [m.s
Radial velocity 2 2 2 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Orbital phase Orbital phase Orbital phase
HARPS HARPS-N HIRES PFS a 10 ] 1 0 [m.s
Radial velocity 10 7400 7600 7800 8000 8200 8400 8600 Time [BJD - 2 450 000] b 10 ] 1 0 [m.s
Radial velocity 10 The case of HIP41378 (III) 8400 8450 8500 8550 8600 • Intensive RVs observations (450+ epochs) by 4Time stabilised [BJD - 2 450 000] spectrographs overHIP41378 4 years b (HARPS, HARPS-N, HIRES, PFS) HIP41378 c HIP41378 g c d e 2 2 2
• Detection] of the Doppler signals of HIP41478 b, c, and f (+ non- 1 transiting 0 planet g) 0 0 [m.s
Radial velocity 2 2 2 • RV signal of planet f perfectly matches with only one solution allowed0.00 by the0.25 photometry0.50 0. 75(the 542-day1.00 0 solution).00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 HIP41378 d HIP41378 e HIP41378 f af g h 102 2 2 ] ] 1 1 0 0 0 [m.s [m.s
Radial velocity 2 2 2 Radial velocity 10 0.740000 0.25 0.5076000.75 1.0078000.00 0.25 80000.50 0.75 82001.00 0.00 08400.25 0.50 0.860075 1.00 Orbital phase TimeOrbital [BJD - phase 2 450 000] Orbital phase b 10 HARPS HARPS-N HIRES PFS ] 1
Santerne et al. (2020) 0 [m.s
Radial velocity 10 8400 8450 8500 8550 8600 Time [BJD - 2 450 000] HIP41378 b HIP41378 c HIP41378 g c d e 2 2 2 ] 1 0 0 0 [m.s
Radial velocity 2 2 2 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00
HIP41378 d HIP41378 e HIP41378 f f g h 2 2 2 ] 1 0 0 0 [m.s
Radial velocity 2 2 2 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Orbital phase Orbital phase Orbital phase
HARPS HARPS-N HIRES PFS HIP41378: a good system for ARIEL (& JWST)
14
12
10
b c g d e f 8
6
J-band magnitude of4 the host star
1011 1010 109 108 107 106 105 104 2 1 Stellar insolation flux [erg.cm .s ]
Santerne et al. (2020) (Not for phase curves) ARIEL needs to know the stellar variability ARIEL needs to know the stellar variability
• Stellar activity is a well known limitation to the interpretation of transmission spectroscopy data ARIEL needs to know the stellar variability
• Stellar activity is a well known limitation to the interpretation of transmission spectroscopy data
• The optical channels of ARIEL will be very important for occulted star spot/faculae
CoRoT-2 ; Bruno et al. (2016) Disentangling spot vs faculae
• To understand the impact of stellar activity on transit data, we need to monitor the host stars over one rotation.
• Impossible to do with ARIEL
HIGH-RESOLUTION SPECTROSCOPY MIGHT HELP
CoRoT-2 ; Bruno et al. (2016) Spectroscopic diagnoses for stellar activity
• Chromospheric emission (CaII H&K, Hα, Na D, etc…)
• Line bisector & width
• Spectropolarimetry (Zeeman effect) Spectroscopic diagnoses for stellar activity
• Chromospheric emission (CaII H&K, Hα, Na D, etc…)
• Line bisector & width
• Spectropolarimetry (Zeeman effect)
Still need to understand the physical link between these spectroscopic diagnoses and the spot / faculae configuration + impact on transmission spectroscopy Spectroscopic diagnoses for stellar activity
• Chromospheric emission (CaII H&K, Hα, Na D, etc…)
• Line bisector & width
• Spectropolarimetry (Zeeman effect)
Still need to understand the physical link between these spectroscopic diagnoses and the spot / faculae configuration + impact on transmission spectroscopy
Stellar activity is a limitation for all of us (RVs, PLATO, ARIEL), we should join our efforts … La campagne d’activité K2: donnéesLa campagne obtenues d’activité sur la K2: M données obtenues sur la M
1.10 Stellar activity campaign HARPS SOPHIE 100 K2 TRAPPIST-N (z’) TRAPPIST-N (V) 1.10 HARPS SOPHIE 1.08 100 K2 TRAPPIST-N (z’) TRAPPIST-N (V)
] 50 1 1.08 1.06
] 50 1 0 1.06 0 1.04 50 1.04 Relative flux 50 1.02 Relative flux
Radial velocity [m.s 100 1.02
Radial velocity [m.s 100 1.00 1.00 150 150 0.98 100 110 120 130 140 150 160 170 0.98 100 110 120 Time130 [BJD - 2 458 000]140 150 160 170 Time [BJD - 2 458 000]
Lopez et al. (in prep.) AnalyseAnalyse préliminaire préliminaire de de la laspectropolarimétrie. spectropolarimétrie. Crédit:Crédit: Jean-FrançoisJean-François Donati Donati36 36 Take-home messages
• RVs are an important input for the ARIEL mission (exoplanet mass, eccentricity, ephemerides)
• We can perform RV preparatory observations on ARIEL secured targets, at least before PLATO’s launch
• Need to perform spectroscopic observations contemporaneous with the ARIEL observations, at least for the most active stars.
• We need to work together to model stellar activity based on spectroscopic diagnoses and correct for it.