An Eccentric Massive Jupiter Orbiting a Sub-Giant on a 9.5 Day Period Discovered in the Transiting Exoplanet Survey Satellite Full Frame Images

TL;DR

This paper reports the discovery of TOI-172 b, a massive, eccentric hot Jupiter orbiting a slightly evolved G-star with a 9.5-day period, identified solely from TESS full frame images, highlighting a new detection method.

Contribution

First confirmed planet from TESS full frame images analysis without targeted two-minute cadence data, demonstrating the capability of TESS FFI data for exoplanet discovery.

Findings

TOI-172 b is a massive hot Jupiter with 5.42 M_J and 0.97 R_J.

The planet orbits its star with an eccentricity of 0.38.

The host star is a slightly evolved G-star with 1.128 M_sun and 1.777 R_sun.

Abstract

We report the discovery of TOI-172 b from the Transiting Exoplanet Survey Satellite (TESS) mission, a massive hot Jupiter transiting a slightly evolved G-star with a 9.48-day orbital period. This is the first planet to be confirmed from analysis of only the TESS full frame images, because the host star was not chosen as a two minute cadence target. From a global analysis of the TESS photometry and follow-up observations carried out by the TESS Follow-up Observing Program Working Group, TOI-172 (TIC 29857954) is a slightly evolved star with an effective temperature of $T_{\rm eff}$ =$5645\pm50$ K, a mass of $M_{\star}$ = $1.128^{+0.065}_{-0.061}$ $M_{\odot}$, radius of $R_{\star}$ = $1.777^{+0.047}_{-0.044}$ $R_{\odot}$, a surface gravity of $\log$ $g_{\star}$ = $3.993^{+0.027}_{-0.028}$, and an age of $7.4^{+1.6}_{-1.5}$ Gyr. Its planetary companion (TOI-172 b) has a radius of $R_{\rm P}$ = $0.965^{+0.032}_{-0.029}$ $R_{\rm J}$, a mass of $M_{\rm P}$ = $5.42^{+0.22}_{-0.20}$ $M_{\rm J}$, and is on an eccentric orbit ($e = 0.3806^{+0.0093}_{-0.0090}$). TOI-172 b is one of the few known massive giant planets on a highly eccentric short-period orbit. Future study of the atmosphere of this planet and its system architecture offer opportunities to understand the formation and evolution of similar systems.

Figures

Tables

Table 1. Table 1: Literature and Measured Properties for TOI-172

Other identifiers
TIC 29857954
TYC 6932-00301-1
2MASS J21063165-2641333
Parameter	Description	Value	Source
${\alpha }_{J2000}$	Right Ascension (RA)	21:06:31.65	1
${\delta }_{J2000}$	Declination (Dec)	-26:41:34.29	1
BT	Tycho BT mag.	12.211 $\pm$ 0.203	2
VT	Tycho VT mag.	11.382 $\pm$ 0.125	2
$G$	Gaia $G$ mag.	11.193$\pm$0.02	3,4
$T$	TESS mag.	10.711$\pm$0.019	4
J	2MASS J mag.	10.135 $\pm$ 0.03	5, 6
H	2MASS H mag.	9.825 $\pm$ 0.03	5, 6
KS	2MASS $K_S$ mag.	9.722 $\pm$ 0.02	5, 6
WISE1	WISE1 mag.	9.673 $\pm$ 0.03	7
WISE2	WISE2 mag.	9.718 $\pm$ 0.03	7
WISE3	WISE3 mag.	9.763 $\pm$ 0.052	7
WISE4	WISE4 mag.	8.529 $\pm$ 0.516	7
${\mu }_{\alpha }$	Gaia DR2 proper motion	-4.711 $\pm$ 0.094	3,4
	in RA (mas yr-1)
${\mu }_{\delta }$	Gaia DR2 proper motion	-54.25 $\pm$ 0.069	3,4
	in DEC (mas yr-1)
$v\sin i_{\star }$	Rotational velocity ( km s-1)	5.1 $\pm$ 0.5	§2.3
$\left[\mathrm{Fe}/H\right]$	Metallicity	0.14 $\pm$ 0.08	§2.3
$T_{\mathrm{eff}}$	Effective Temperature (K)	5640 $\pm$ 50	§2.3
$\log g_{\star }$	Surface Gravity (cgs)	3.97 $\pm$ 0.1	§2.3
$\pi$	Gaia Parallax (mas)	2.972 $\pm$ 0.06†	3,4
$RV$	Systemic radial	$-6.247\pm 0.081$	§2.3
	velocity ( km s-1)
$d$	Distance (pc)	$336.47\pm {6.79}^{†}$	3,4
$U^{\ast }$	Space Velocity ( km s-1)	$26.24\pm 0.46$	§2.7
$V$	Space Velocity ( km s-1)	$-71.52\pm 1.68$	§2.7
$W$	Space Velocity ( km s-1)	$-1.31\pm 0.27$	§2.7

Table 2. Table 2: Relative Radial Velocities for TOI-172

${\mathrm{BJD}}_{\mathrm{TDB}}$	RV (m s-1)	${\sigma }_{RV}$ (m s-1)	Bisectors	Instrument
2458410.551705	-5716.4	7.4	-1.0$\pm$10.0	FEROS
2458411.636884	-5739.9	9.9	6.0$\pm$13.0	FEROS
2458415.573206	-6562.4	8.6	-4.0$\pm$11.0	FEROS
2458418.552974	-6100.9	7.1	-18.0$\pm$10.0	FEROS
2458419.511487	-5865.5	7.9	38.0$\pm$11.0	FEROS
2458423.625864	-6654.1	8.3	-15.0$\pm$11.0	FEROS
2458424.531225	-6584.4	7.0	8.0$\pm$10.0	FEROS
2458425.597896	-6474.7	7.8	-7.0$\pm$11.0	FEROS
2458427.583262	-6149.8	7.4	-8.0$\pm$10.0	FEROS
2458429.542763	-5682.3	7.4	13.0$\pm$10.0	FEROS
2458430.614729	-5660.9	17.1	35.0$\pm$19.0	FEROS
2458376.730770	-593.6	27.0	13.3$\pm$19.0	TRES
2458390.710107	122.0	17.6	7.5$\pm$14.7	TRES
2458397.687254	-359.0	14.0	23.7$\pm$21.2	TRES
2458398.649461	-206.5	26.2	16.5$\pm$18.4	TRES
2458400.688662	230.9	22.4	8.5$\pm$21.9	TRES
2458409.670769	78.7	35.8	-56.7$\pm$33.4	TRES
2458413.668975	-569.8	43.0	-123.3$\pm$51.7	TRES
2458415.631951	-499.7	21.2	-5.3$\pm$22.6	TRES
2458416.639897	-373.8	30.9	13.6$\pm$29.0	TRES
2458417.601186	-194.6	23.1	1.0$\pm$21.6	TRES
2458418.659549	-8.6	24.7	14.1$\pm$28.5	TRES
2458419.685623	274.1	23.1	-10.4$\pm$35.5	TRES
2458420.614450	422.1	20.0	16.4$\pm$16.1	TRES
2458423.681839	-559.6	46.0	-9.8$\pm$23.3	TRES
2458424.653259	-520.3	26.9	25.1$\pm$25.3	TRES
2458426.597813	-250.4	22.6	10.5$\pm$25.6	TRES
2458428.594704	126.1	16.5	29.0$\pm$15.2	TRES
2458430.627687	291.2	19.2	37.342$\pm$2.3	TRES
2458438.602799	228.7	16.3	25.6$\pm$11.2	TRES
2458439.594972	403.1	16.6	-3.7$\pm$11.5	TRES
2458440.586306	-29.0	12.8	1.5$\pm$13.9	TRES
2458441.577056	-528.6	15.8	30.8$\pm$13.1	TRES
2458442.582383	-617.8	13.8	21.4$\pm$14.7	TRES
2458443.586964	-548.2	19.5	-15.5$\pm$22.6	TRES
2458444.630963	-471.7	39.9	-67.3$\pm$45.2	TRES
2458446.575996	-135.6	18.0	-33.4$\pm$18.1	TRES
2458447.572304	104.0	15.1	29.4$\pm$15.9	TRES

Table 3. Table 3: Median values and 68% confidence interval for global model of TOI-172

Parameter	Units	Values
Stellar Parameters:
$M_{\ast }$	Mass ($M_{☉}$)	${1.128}_{-0.061}^{+0.065}$
$R_{\ast }$	Radius ($R_{☉}$)	${1.777}_{-0.044}^{+0.047}$
$L_{\ast }$	Luminosity ($L_{☉}$)	${2.89}_{-0.18}^{+0.19}$
${\rho }_{\ast }$	Density (cgs)	${0.286}_{-0.023}^{+0.022}$
$\log g$	Surface gravity (cgs)	${3.993}_{-0.028}^{+0.027}$
$T_{\mathrm{eff}}$	Effective Temperature (K)	$5645\pm 50$
$\left[\mathrm{Fe}/H\right]$	Metallicity (dex)	${0.148}_{-0.080}^{+0.079}$
${\left[\mathrm{Fe}/H\right]}_0^{†}$	Initial Metallicity	${0.172}_{-0.078}^{+0.074}$
$Age$	Age (Gyr)	${7.4}_{-1.5}^{+1.6}$
$EE{P}^{‡}$	Equal Evolutionary Point	${456.0}_{-6.8}^{+3.5}$

Table 4. Table 4: Median values and 68% confidence interval for global model of TOI-172

Parameter	Description (Units)	Values
$P$	Period (days)	${9.47725}_{-0.00079}^{+0.00064}$
$R_P$	Radius ($R_J$)	${0.965}_{-0.029}^{+0.032}$
$T_C$	Time of conjunction (${\mathrm{BJD}}_{\mathrm{TDB}}$)	$2458326.9190\pm 0.0017$
$T_0^{†}$	Optimal conjunction Time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	$2458345.8734\pm 0.0013$
$a$	Semi-major axis (AU)	$0.0914\pm 0.0017$
$i$	Inclination (Degrees)	${88.2}_{-1.0}^{+1.1}$
$e$	Eccentricity	${0.3806}_{-0.0090}^{+0.0093}$
${\omega }_{\ast }$	Argument of Periastron (Degrees)	$57.1\pm 1.7$
$T_{\mathrm{eq}}$	Equilibrium temperature (K)	${1198}_{-17}^{+18}$
$M_P$	Mass ($M_J$)	${5.42}_{-0.20}^{+0.22}$
$K$	RV semi-amplitude (m/s)	$517.6\pm 6.2$
$logK$	Log of RV semi-amplitude	$2.7140\pm 0.0052$
$R_P/R_{\ast }$	Radius of planet in stellar radii	${0.05588}_{-0.00092}^{+0.00091}$
$a/R_{\ast }$	Semi-major axis in stellar radii	${11.09}_{-0.30}^{+0.28}$
$\delta$	Transit depth (fraction)	$0.00312\pm 0.00010$
$Depth$	Flux decrement at mid transit	$0.00312\pm 0.00010$
$\tau$	Ingress/egress transit duration (days)	${0.01093}_{-0.00050}^{+0.00085}$
$T_{14}$	Total transit duration (days)	${0.1964}_{-0.0029}^{+0.0028}$
$T_{\mathrm{FWHM}}$	FWHM transit duration (days)	$0.1853\pm 0.0029$
$b$	Transit Impact parameter	${0.22}_{-0.14}^{+0.12}$
$b_S$	Eclipse impact parameter	${0.43}_{-0.27}^{+0.23}$
${\tau }_S$	Ingress/egress eclipse duration (days)	${0.0229}_{-0.0024}^{+0.0054}$
$T_{S,14}$	Total eclipse duration (days)	${0.355}_{-0.043}^{+0.026}$
$T_{S,\mathrm{FWHM}}$	FWHM eclipse duration (days)	${0.332}_{-0.048}^{+0.027}$
${\delta }_{S,3.6\mu m}$	Blackbody eclipse depth at 3.6$\mu$m (ppm)	${115.2}_{-6.3}^{+7.8}$
${\delta }_{S,4.5\mu m}$	Blackbody eclipse depth at 4.5$\mu$m (ppm)	${176.7}_{-8.6}^{+10.}$
${\rho }_P$	Density (cgs)	${7.53}_{-0.72}^{+0.65}$
$log{g}_P$	Surface gravity	${4.162}_{-0.031}^{+0.026}$
$\Theta$	Safronov Number	${0.908}_{-0.031}^{+0.030}$
$⟨F⟩$	Incident Flux (109 erg s-1 cm-2)	${0.407}_{-0.022}^{+0.026}$
$T_P$	Time of Periastron (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2458326.549}_{-0.027}^{+0.025}$
$T_S$	Time of eclipse (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2458323.485}_{-0.049}^{+0.052}$
$T_A$	Time of Ascending Node (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2458325.863}_{-0.024}^{+0.025}$
$T_D$	Time of Descending Node (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2458328.667}_{-0.049}^{+0.050}$
$ecos{\omega }_{\ast }$		${0.2065}_{-0.0082}^{+0.0086}$
$esin{\omega }_{\ast }$		$0.319\pm 0.012$
$M_P\sin i$	Minimum mass ($M_J$)	${5.42}_{-0.21}^{+0.22}$
$M_P/M_{\ast }$	Mass ratio	${0.004587}_{-0.00011}^{+0.000100}$
$d/R_{\ast }$	Separation at mid transit	${7.17}_{-0.23}^{+0.24}$
$P_T$	A priori non-grazing transit prob	${0.1316}_{-0.0043}^{+0.0044}$
$P_{T,G}$	A priori transit prob	${0.1472}_{-0.0048}^{+0.0050}$
$P_S$	A priori non-grazing eclipse prob	${0.0677}_{-0.0016}^{+0.0020}$
$P_{S,G}$	A priori eclipse prob	${0.0757}_{-0.0018}^{+0.0023}$
Wavelength Parameters:		i’	z’	TESS
$u_1$	linear limb-darkening coeff	$0.325\pm 0.051$	$0.263\pm 0.050$	$0.310\pm 0.049$
$u_2$	quadratic limb-darkening coeff	${0.271}_{-0.051}^{+0.050}$	${0.270}_{-0.050}^{+0.049}$	${0.263}_{-0.050}^{+0.049}$
$A_D$	Dilution from neighboring stars	${0.0017734}_{-0.0000062}^{+0.0000063}$	$0.014331\pm 0.000051$	$0.007830\pm 0.000028$
Telescope Parameters:		FEROS	TRES
${\gamma }_{\mathrm{rel}}$	Relative RV Offset (m/s)	$-{6240.2}_{-9.5}^{+10.}$	$-195.7\pm 4.2$
${\sigma }_J$	RV Jitter (m/s)	${30.0}_{-8.4}^{+12}$	${6.4}_{-6.4}^{+7.7}$
${\sigma }_J^2$	RV Jitter Variance	${900}_{-430}^{+900}$	${41}_{-95}^{+160}$
Transit Parameters:		LCO UT 2018-09-22 (i’)	LCO UT 2018-10-11 (z’)	TESS
${\sigma }^2$	Added Variance	${0.0000192}_{-0.0000025}^{+0.0000030}$	${0.0000085}_{-0.0000013}^{+0.0000016}$	$-{0.0000000139}_{-0.0000000078}^{+0.0000000083}$
$F_0$	Baseline flux	$1.00051\pm 0.00046$	${0.99970}_{-0.00037}^{+0.00038}$	$1.000007\pm 0.000013$

Full text

planetary systems, planets and satellites: detection, stars: individual (TOI-172)

††software: EXOFASTv2 (Eastman et al. 2013; Eastman 2017), AstroImageJ (Collins et al. 2017)††facilities: TESS, FLWO 1.5m (Tillinghast Reflector Echelle Spectrograph), 4.1-m Southern Astrophysical Research (SOAR), LCO 0.4m, LCO 1.0m, 2.2m telescope La Silla (Fiber-fed Extended Range Optical Spectrograph)

1 Introduction

In only three decades, the field of exoplanets has rapidly expanded from its infancy to one of the largest and fastest research areas in astrophysics. This is largely due to the success of both ground-based and space-based efforts to discover new planets using the transit and radial velocity (RV) techniques. With the confirmation of thousands of new planets and the identification of a few thousand more candidates, no survey has been more influential to the field than the Kepler mission (Borucki et al. 2010). As the Kepler and re-purposed K2 (Howell et al. 2014) missions have completed, we are now entering the next major chapter in the field of exoplanets with the recent launch of the Transiting Exoplanet Survey Satellite (TESS, Ricker et al. 2015).

Interestingly, we are still attempting to understand one of the first types of planets ever discovered, hot Jupiters. It is commonly believed that close-in giant planets formed farther out in the protoplanetary disk and, through various mechanisms, migrated inward. These highly irradiated, Jovian-sized planets orbit with periods $\leq 10$ days, and typically do not have nearby planetary companions (Steffen et al. 2012; Huang et al. 2016), suggesting that they might disrupt planet formation and the orbits of any existing inner planets as they move inward. However, the discovery of two small planets bracketing the known hot Jupiter, WASP-47b (Becker et al. 2015), suggests that some giant planets can migrate in a dynamically quiet manner or even form in-situ (Huang et al. 2016; Batygin et al. 2016). It has been found that giant planets discovered in more distant orbits tend to have companions (e.g. Knutson et al. 2014; Huang et al. 2016). This possibly supports the idea that their longer orbit allows them to form alongside smaller planets in different parts of the inner disk.

If planetary migration occurs through the gas disk, it must take place during the first $\sim$10 Myr while the gas is still around, and is expected to result in low-eccentricity orbits (Haisch et al. 2001; D’Angelo et al. 2003). However, migration may commonly be influenced by gravitational interactions with other planets or stars. These interactions can increase the planet’s orbital eccentricity (known as “High Eccentricity Migration” (HEM)) and lead to tidal interactions at close approach to the host star that shrink and circularize the orbit (Rasio & Ford 1996; Wu & Murray 2003; Fabrycky & Tremaine 2007; Nagasawa & Ida 2011; Wu & Lithwick 2011). For a Jupiter analogue orbiting a Sun-like star on a period of 0.5 to 10 days, the circularization timescale can range from a few million years to over a hundred billion years depending on semi-major axis (see equation 2 in Adams & Laughlin 2006). Therefore, only long period hot Jupiters (5–10 days) would retain any primordial eccentricity if HEM is the underlying mechanism because they would not have had enough time to circularize. This class of “dynamically young” giant planets, for which the circularization timescales are longer than the system’s current age (also referred to as ”tropical Jupiters”; Yu et al. 2018), offers an opportunity to gain insight into the mechanisms governing hot Jupiter evolution. Previous studies have tried to place constraints on hot Jupiter migration mechanisms by analyzing the eccentricities and orbital architectures of these systems. For example, the orbits of dynamically young hot Jupiters tend to be more eccentric on average, as would be expected if at least a fraction of them have undergone eccentric migration (Quinn et al. 2014; Bonomo et al. 2017). At the same time, the paucity of highly eccentric migrating Jupiters places an upper limit on the prevalence of HEM in the production of these systems (Dawson et al. 2015). The presence of additional giant planets exterior to hot Jupiters but inside the ice line is hard to reconcile with migration via HEM (Schlaufman & Winn 2016), though trends with host star metallicity hint that disk migration and subsequent planet-planet scattering could account for much of the hot Jupiter population (Dawson & Murray-Clay 2013). Indeed, it appears that no single migration channel can produce the known population; a recent review of the relevant literature suggests that the combination of two such mechanisms might be able to explain the observations (Dawson & Johnson 2018). Additional study of the dynamically young planets—and their orbital architectures—can refine our understanding of how these migration mechanisms work together to produce the population of giant planets that we observe.

In this paper, we present the discovery in TESS full frame images leading to follow-up photometry, and precision radial velocity measurements of a dynamically young, massive Jupiter in a $\sim$9.5-day eccentric orbit (0.38) around a sub-giant. Additionally, the evolutionary state of TOI-172 provides a reliable age. The paper is organized in the following way. We present all available observations of TOI-172 in §2 (Table 1 presents the available information on TOI-172 from the literature). Our global analysis of all available observations using EXOFASTv2 is described in §3. We discuss TOI-172 b in the context of all known planets in §4, presenting prospects on future follow-up. We summarize our conclusions in §5.

2 Observations and Archival Data

2.1 TESS Photometry

TOI-172 fell on CCD 4 of Camera 1 of the TESS spacecraft during its first sector of observations (2018 July 25 – August 22), but it was not pre-selected for two-minute cadence observations. After the data were downloaded from the spacecraft, we processed the calibrated 30-minute cadence full frame images (Jenkins et al. 2016) with the MIT Quick Look Pipeline (QLP, C. Huang et al., in preparation). The QLP is a lightweight tool for rapidly producing light curves and identifying transits in all stars observed by TESS, not just those selected for two-minute cadence observations. The QLP extracts photometry by summing the flux within moving circular apertures (following Huang et al. 2015), after using the nebulosity filter111developed by Irwin (2010):http://www.ukirt.hawaii.edu/publications/newsletter/ukirtnewsletter2010spring.pdf to remove scattered background light from the images). After producing light curves, the QLP searches for transits by calculating a Box-Least-Squares periodogram (BLS, Kovács et al. 2002), implementing high-pass filtering and BLS period spacing following Vanderburg et al. (2016). We detected a single repeating transit signal around TOI-172 with a period of 9.48 days, a duration of 4.71 hours, and a flat-bottomed shape (see Figure 1). We notified the community of the discovery via the MIT TESS Alerts portal222https://tess.mit.edu/alerts/ (Ricker & Vanderspek 2018).

Upon the public release of the processed Sector 1 full frame images, we attempted to improve the light curve by extracting photometry from a variety of differently shaped stationary photometric apertures. After some experimentation, and using archival images from the ESO/SERC Southern Sky Atlas (SERC-J; taken in 1975) and the Anglo-Australian Observatory Second Epoch Survey (AAO-SES; 1993) to check for any additional stars nearby, we settled upon the irregularly shaped aperture shown in Figure 2. The light curve extracted from this aperture balanced high photometric precision with minimal contamination from a nearby 12th magnitude star (TIC 29857959) and minimal systematics related to TESS’s “momentum dumps”, when the spacecraft thrusters are fired to reorient the spacecraft and allow the reaction wheels to be spun down. We compared the transit depths from the light curve extracted with the QLP and our simple aperture photometry method, and found consistent results. We proceeded in our analysis using the light curve produced with simple aperture photometry, as it had slightly better photometric precision. We manually removed 8 hours of data (BJDTDB- 2457000 = 1338.4125 to 1338.0792) contaminated by an asteroid passing through the photometric aperture, and we clipped 4$\sigma$ outliers from the light curve (see Figure 1). The corresponding TESS light curve was flattened by using a spline fit with breakpoints every 0.5 days to divide out the best-fit stellar variability (Vanderburg & Johnson 2014).

2.2 Ground-based Photometry from the TESS Follow-up Observing Program Working Group

To rule out false positives, better constrain the ephemeris of TOI-172 b, and measure the depth of the transit, we obtained two photometric transit follow-up observations using the Las Cumbres Observatory (LCO) telescope network (Brown et al. 2013)333https://lco.global/. To predict the next possible transit events for TOI-172 that were observable, we used the TAPIR software package (Jensen 2013). We used the AstroImageJ astronomical observation analysis software to reduce all follow-up photometric observations and perform aperture photometry to extract the light curves. On UT 2018 September 22, we observed the transit of TOI-172 b in the SDSS i*′* filter using the 0.4m LCO telescope located at the Cerro Tololo Inter-American Observatory (CTIO) in Chile. The 0.4m telescopes are equipped with SBIG STX6303 cameras that have a $19\arcmin\times 29\arcmin$ field-of-view, and a 0.57$\arcsec$ pixel scale. On UT 2018 October 11, we observed the transit of TOI-172 b in the z*′* filter on the 1.0m telescope at the McDonald Observatory in Fort Davis, Texas. The 1.0m telescope has a Sinistro camera with a $16.5\arcmin\times 26.5\arcmin$ FOV and a pixel scale of 0.389$\arcsec$ pixel*-1*. In each case, an ingress of the transit of TOI-172 b was observed on the target star. In both observations, only an ingress was observable and the exposure time was 50s. These observations are consistent with what was observed by TESS (see Figure 3). Therefore, the fading events are localized to within 15$\arcsec$ of TOI-172.

2.3 TRES Spectroscopy

Spectra of TOI-172 were obtained on 27 occasions with a resolving power of R$\sim$44000 using the Tillinghast Reflector Echelle Spectrograph (TRES; Fűrész 2008)444http://www.sao.arizona.edu/html/FLWO/60/TRES/GABORthesis.pdf mounted on the 1.5m Tillinghast Reflector at the Fred L. Whipple Observatory (FLWO) on Mt. Hopkins, AZ. For a description of the reduction and radial velocity (RV) extraction pipeline, see Buchhave et al. (2010). Our procedure differed only in the generation of the template used for cross-correlation. We derived relative RVs by cross-correlating against the strongest spectrum, and we shifted and median-combined the spectra to produce a high-SNR template spectrum. We then cross-correlated each observed spectrum against that template to produce our final relative RVs, which are given in Table 2 and shown in Figure 4. Bisector spans were calculated for the TRES RVs using the technique described in Torres et al. (2007). There are no correlations between the bisector spans and the measured RV values and no scatter in the bisectors beyond their uncertainties (which are small compared to the RV variation), supporting the premise that TOI-172 is being periodically transited or eclipsed. We also derive the absolute RVs via cross-correlation against synthetic templates created using Kurucz model atmospheres (Kurucz 1992). We calculate the instrumental zero-point through nightly monitoring of RV standards, which we place on the absolute RV scale of Nidever et al. (2002). Using these observations, we determine the absolute center-of-mass velocity of TOI-172 to be -6.247$\pm$0.081  km s*-1*(consistent with the absolute RV from Gaia DR2 of -5.89 $\pm$ 0.67).

To determine the stellar parameters of TOI-172, we analyzed the TRES spectra using the Stellar Parameter Classification (SPC) analysis package (Buchhave et al. 2012). From this analysis, we estimated the effective temperature, metallicity, surface gravity, and rotational velocity of TOI-172 to be: $T_{\rm eff}$ = 5640 $\pm$ 50 K, $\log{g_{\star}}$ = 3.97 $\pm$ 0.1, [$m/H$] = 0.14 $\pm$ 0.08, and $v\sin{I_{*}}$ = 5.1 $\pm$ 0.5 km s*-1*. We use the $T_{\rm eff}$ and $\left[{\rm Fe}/{\rm H}\right]$ as a prior in the EXOFASTv2 global fit (see §3).

2.4 FEROS Spectroscopy

We also obtained 9 R=48000 spectra of TOI-172 between UT 2018 October 19 and November 5 using the FEROS spectrograph (Kaufer et al. 1999) mounted on the 2.2m MPG telescope at La Silla observatory in Chile. Each spectrum achieved a signal-to-noise ratio of $\sim 60$–$100$ per spectral resolution element with exposure times of 600 sec. The instrumental drift was determined via comparison with a simultaneous fiber illuminated with a ThAr+Ne lamp. The data were processed with the CERES suite of echelle pipelines (Brahm et al. 2017), which produce radial velocities and bisector spans in addition to reduced spectra.

2.5 High Resolution Imaging

The relatively large 21″ pixels of TESS can result in photometric contamination from nearby sources. These must be accounted for to help rule out astrophysical false positives, such as background eclipsing binaries, and to correct the estimated planetary radius, initially derived from the diluted transit in a blended light curve (Ciardi et al. 2015; Ziegler et al. 2018). We searched for close companions to TOI-172 with speckle imaging on the 4.1-m Southern Astrophysical Research (SOAR) telescope (Tokovinin 2018) on UT 2018 September 25, and again in better conditions on UT 2018 October 21. We also obtained adaptive optics (AO) images of the target on UT 2018 November 14 using Gemini/NIRI. For these observations, 9 science frames with exposure time 11 seconds each were collected, with the telescope dithered between each frame. For a subset of the frames, the raw data showed signs of stripping, and so we discarded these frames and combined only the 6 good frames for the analysis. We flat field and sky subtract the frames, using a sky background constructed by median combining the dithered images, and then align and combine the images. Figure 5 shows the $5\sigma$ detection limits along with the AO image and speckle auto-correlation function.

A nearby star was detected in both the speckle and the AO observations. The object is measured at a separation of $1.104\arcsec$ and an I-band contrast of 4.9 mag in the speckle images, and at a separation of $1.099\arcsec$ and a Br$\gamma$ contrast of 4.5 mag in the AO images. This would result in a 371 au projected separation if the companion is at the same distance as TOI-172. To test this assumption, we use the broadband photometry in these two bands for TOI-172 (2MASS $K_{s}$ = 9.722 and USNO $I^{\prime}$ = 10.50, Monet et al. 2003) and the measured contrasts for the companion ($\Delta$Br$\gamma$ = 4.5 and $\Delta$I = 4.9) to derive the $I-K_{s}$ colors to be 0.78 and 1.18, respectively. Using these colors and the MIST stellar evolution models (Dotter 2016; Choi et al. 2016; Paxton et al. 2011, 2013, 2015) at a log(age) = 9.9 and a solar metallicity (the nearest isochrone grid for TOI-172), we estimate a photometric distance to TOI-172 to be 318 pc. This is in close agreement with the Gaia distance (336.47 pc). However, assuming the companion is on the main sequence, we estimate its distance to be 1092 pc. When using other MIST isochrone grids near the one adopted here, we only see a small change in the derived photometric distances, not nearly enough to explain the large difference measured between TOI-172 and the visual companion. This discrepancy suggests that the visual companion is likely a background object, and not gravitationally bound to the planet host. More data are required to confirm this conclusion, either in the form of more photometry to further characterize the SED of the visual companion, or additional astrometric measurements that confirm whether the two stars share common proper motion.

The nearby star is not in Gaia DR2 or the TESS input catalog, and consequently was not accounted for in the contamination correction for TOI-172. It would take a 28.5% deep eclipse of the nearby faint companion to cause the blended depth seen in our aperture for TOI-172. The high contrast between the two stars significantly reduces the possibility that the nearby star is a background eclipsing binary resulting in a false positive planetary transit signal, as does subsequent radial velocity follow-up. We know that such a faint companion is unable to significantly affect the RVs of TOI-172 because its contribution to the line profile is so small. While it is true that even a faint companion could affect the RVs slightly (even if it is below the noise level of the CCF), this would only be at the level of m/s, not hundreds of m/s (TOI-172b k = 517  m s*-1*, see the analysis of blended CCFs in Buchhave et al. 2011). Therefore, the spectroscopy proves that the planetary companion orbits our target. Assuming the primary star is the planet host, the additional flux from the nearby star results in only a negligible correction upwards to the initially derived planet radius ($\sim$0.5%). We account for the blending from this nearby companion in our global fit (see §3).

2.6 SED Analysis

Due to the presence of a nearby visual companion (see Figure 5), we are unable to simultaneously fit the Spectral Energy Distribution (SED) within the EXOFASTv2 global analysis. Instead, we fit the combined SED of the two stars separately from the join transit and RV analysis. The companion is blended in each of the broadband photometric observations. From our analysis of the speckle high-resolution imaging, we know that the nearby companion has an I-band contrast of 4.9 mag and a Br$\gamma$ contrast of 4.5 mag. Using the available photometric observations (see Table 1), we fit the broadband SED of TOI-172 spanning 0.2–20 $\mu$m (Figure 6). Assuming both stars have the same $A_{V}$, we use the $\Delta$I and $\Delta$Br$\gamma$ contrasts to fit an SED to the nearby companion. Each flux measurement is fit using the stellar atmosphere models of Kurucz (1992). The distance for TOI-172 is adopted from the measured Gaia parallax and we use the SPC determined $T_{\rm eff}$, $\log{g_{\star}}$, and $\left[{\rm Fe}/{\rm H}\right]$ as Gaussian priors on the fit. The only free parameter is the extinction ($A_{V}$) which is constrained at its upper bound by the maximum permitted line-of-sight extinction from Schlegel et al. (1998). Our final best-fit SED for TOI-172 has a reduced $\chi^{2}$ of 1.7 and an extinction $A_{V}=0.08\pm 0.04$, and is shown in Figure 6. We integrated the best-fit SED to determine the unextincted bolometric flux (correcting for the contamination of the companion) received at Earth, $F_{\rm bol}=8.16\pm 0.19\times 10^{-10}$ erg s*-1* cm*-2*. Using the Gaia parallax (corrected for the systematic offset reported by Stassun & Torres 2018) combined with the adopted $T_{\rm eff}$ from this analysis, we are able to measure the radius of TOI-172 to be $R_{\star}=1.787\pm 0.049$ $\,R_{\sun}$, after accounting for the presence of the nearby companion seen in our high-resolution imaging. We use this determined $R_{\star}$ as a prior for the EXOFASTv2 global analysis (see §3). Using our two-component SED fit, we determine the flux contribution of the nearby companion to be 0.91% (TESS), 0.46% (g*′), and 1.07% (z′*). We note that the contribution from the companion would correspond to a change in the measured TESS transit depth $<$1$\sigma$.

2.7 Location in the Galaxy, UVW Space Motion, and Galactic Population

TOI-172 is located at $\alpha_{\rm J2000}=21^{h}06^{m}31\fs 65$ and $\delta_{\rm J2000}=-26\arcdeg 41\arcmin 34\farcs 29$, and from Gaia DR2 the parallax is $2.89\pm 0.06$ mas (applying the correction from Stassun & Torres 2018), corresponding to a distance of $336.47\pm 6.79$ pc ignoring the Lutz-Kelker bias, which can cause measured parallaxes to be larger than they are due to the assumption that the number of observable stars increases as you go farther out (Lutz & Kelker 1973). This results in TOI-172 being 217.6 pc below the Galactic plane. Combining the Gaia DR2 proper motions of $(\mu_{\alpha},\mu_{\delta})=(-4.711\pm 0.094,-54.25\pm 0.069)~{}{\rm mas~{}yr}^{-1}$, the Gaia parallax, and the absolute radial velocity as determined from the TRES spectroscopy of $-6.25\pm 0.081~{}{\rm km~{}s^{-1}}$, we determine the three-dimensional Galactic space motion of $(U,V,W)=(26.24\pm 0.46,-71.52\pm 1.68,-1.31\pm 0.27)~{}{\rm km~{}s^{-1}}$, where positive $U$ is in the direction of the Galactic center. We adopt the Coşkunoǧlu et al. (2011) determination of the solar motion with respect to the local standard of rest. The large asymmetric drift (large negative V velocity) of the host star, combined with its relatively large vertical height below the plane, suggests that the star could potentially be a member of the thick disk. Indeed, TOI-172 has a 43.9% chance of being in the thin disk according to the classification scheme of Bensby et al. (2003). However, this conclusion is somewhat contraindicated by the slightly super-solar metallicity of the host star. We suggest a measurement of the star’s detailed elemental abundances (in particular [$\alpha$/Fe]) could clarify the Galactic population to which this star belongs.

3 EXOFASTv2 Global Fit for TOI-172

We use the EXOFASTv2 modeling suite (Eastman et al. 2013; Eastman 2017) to perform a simultaneous fit of the available photometric and spectroscopic observations to gain a full understanding of the TOI-172 system. EXOFASTv2 is heavily based on the original EXOFAST modeling suite (Eastman et al. 2013) but provides flexibility in allowing the user to simultaneously fit the SED, RV observations from multiple instruments, and an arbitrary number of planets. We simultaneously fit the full frame image TESS light curve (see §2.1 and Figure 1), accounting for the effect of the 30-minute cadence smearing on the light curve, the follow-up ingresses observed by LCO, and the radial velocity observations from TRES and FEROS (see Figure 4). From our Speckle observations and two-component SED analysis, we found that the nearby companion 1.1$\arcsec$ from TOI-172 contributes 0.91%, 0.46%, and 1.07% of the total flux of the system in the TESS, g*′, and z′* band-passes. To properly deblend the TESS and follow-up observations from the previously unknown companion, we include these flux contribution with a 5% error as Gaussian priors in the EXOFASTv2 global fit. This error has no influence on the determined results.

Because accurate TESS pixel response function (PRF) models are not yet available, we did not attempt to deblend the TESS light curve from contaminating flux from TIC 29857959, the 12th magnitude star 75 arcseconds northwest of TOI-172 (see the discussion of this object in Section 2.1 and Figure 2). We did, however, confirm that the neighbor’s contaminating flux does not significantly dilute the transit depth of TOI-172 b using several methods. First, we extracted the light curve of TOI-172 from even smaller apertures than the one shown in Figure 2, and found that decreasing the aperture size had no effect on the depths of the transits (empirically showing dilution is not an important factor). We also estimated the local TESS PRF by examining TESS images of the nearby isolated bright star (TIC 29857846). Inspection of these images showed that in this region of the TESS field of view, about 75% of the total flux falls within one pixel of the peak of the PRF, and virtually all of the flux falls within about 6 pixels of the peak. The photometric aperture for TOI-172 covers about 10% of the detector area within six pixels of TIC 29857959, so only about 2.5% of the neighboring star’s total flux contaminates TOI-172’s aperture. Since TIC 29857959 is about 1.3 magnitudes fainter than TOI-172, the contamination from TIC 29857959 should only be about 1% the total flux in the aperture, much smaller than the uncertainties on the depth of the transit, and therefore negligible for the transit fitting555The uncertainty on the transit depth is about 4%, so 1% dilution affects the measured depth by much less than 1$\sigma$.. Finally, we note that this estimate of about 1% contamination from TIC 29857959 is consistent with the contamination estimated by version 7 of the TESS Input Catalog using pre-launch estimates of the PRF.

To characterize the host star within the fit, we use the MESA Isochrones and Stellar Tracks (MIST) stellar evolution models (Dotter 2016; Choi et al. 2016; Paxton et al. 2011, 2013, 2015). We enforce Gaussian priors on $T_{\rm eff}$ (5640$\pm$50 K) and $\left[{\rm Fe}/{\rm H}\right]$ (0.14$\pm$0.08) from the SPC analysis of the TRES spectra (see §2.3). We also place a Gaussian prior on $R_{*}$ of 1.783$\pm$0.049 $\,R_{\sun}$ from the two-component SED analysis that included the Gaia DR2 parallax (see §2.6). The final determined system parameters for TOI-172 are shown in Tables 3 & 4. Our determined R*⋆* is larger than what was listed in the TESS input catalog (TIC) because version 7 of the TIC did not have a Gaia parallax for TOI-172 and relied on color relations that are unable to distinguish between dwarfs and sub giants.

4 Discussion

Our global analysis indicates that TOI-172 has interesting characteristics that warrant further study. Specifically, TOI-172 b is now one of only now four known planets that has a highly eccentric orbit ($>0.3$), a high planetary mass ($>$3 $\,M_{\rm J}$), relatively short period ($<$20 days), and is bright enough ($V<$12) to be well suited for atmospheric characterization666https://exoplanetarchive.ipac.caltech.edu/; Akeson et al. (2013). The host star has a mass of $M_{\star}$ = $1.128^{+0.065}_{-0.061}$ $M_{\odot}$, a radius of $R_{\star}$ = $1.777^{+0.047}_{-0.044}$, a surface gravity of $\log$ $g_{\star}$ = $3.993^{+0.027}_{-0.028}$, and an age of $7.4^{+1.6}_{-1.5}$ Gyr. Therefore, TOI-172 appears to have just evolved off the main sequence and to be entering into the relatively short sub-giant phase (see Figure 7).

4.1 Tidal Evolution and Irradiation History

To gain a better understanding of the past and future evolution of TOI-172 b’s orbit, we use the latest version of POET777https://github.com/kpenev/poet, where the results of our EXOFASTv2 global analysis (see §3) are used as boundary conditions. POET is a tool for calculating the evolution of a planetary orbit (circular-aligned) as a result of tidal dissipation (see Penev et al. 2014 for a detailed description of the original version of POET). Here we present an overview of the major changes that were used for the analysis of TOI-172 b. The current version allows for inclined and eccentric orbits, where either object in the binary system can be a star or a planet. For the purposes of TOI-172, the difference between a star and a planet is that stars evolve (e.g. their radius changes) while planets do not. We assume that the star follows the MIST evolutionary tracks used in the EXOFASTv2 global model and that the rotation period of TOI-172 is always longer than the orbital period of the planet. We note that this is not strictly true, since stars similar to TOI-172 typically have a rotation period less than $\sim$9.5 days earlier in their lifetime while they are on the main-sequence. This assumption only affects the very early part of the analysis (near a zero-age main sequence), since the part of the evolution after the star has started spinning slower than the orbit is determined entirely by the present state of the system. We note that the estimated $v\sin{I_{*}}$ from the TRES spectroscopy suggests a maximum rotation period of 17.6 days.

For TOI-172, orbital evolutionary tracks were calculated for Q*⋆= 106, 107 and 108, and for each of those, QP* = 106, 107 and 108 (see Figure 8). The tidal quality factor (Q) defines the efficiency of tidal dissipation within the planet or star. Each track uses initial conditions that reproduce the present day orbital period and eccentricity of the system. Unfortunately, due to the high density of the planet and its relatively large semi-major axis, we are unable to produce meaningful constraints on QP or Q*⋆* (see Figure 8). In particular, even for Q*⋆=106 and QP*=106 the amount of circularization this system has undergone is relatively low. When using Q*⋆* = 105, we are not able to find an initial eccentricity large enough to replicate the present eccentricity observed for TOI-172 b. However, we are unable to try initial eccentricities larger than about e=0.6, because the Taylor series expansion of the tidal potential in eccentricity diverges past that point.

This system contradicts normal conventional wisdom that tidal circularization is dominated by tides raised on the planet since the rate of circularization scales as (M’/M)$\times$R5, where M and R are the mass and radius of the body experiencing the tides, and M’ is the mass of the companion (Adams & Laughlin 2006). For a typical Jupiter mass planet around a Solar-type star, the contribution from tides raised on the planet is stronger than that from tides raised on the star by a factor of ten. However, TOI-172 b is more massive than Jupiter (MP = 5.4 $\,M_{\rm J}$) and the host star is larger than the Sun (R*⋆* = 1.78 R*⊙*). Compared to the fiducial case, these contribute to an increase by a factor of nearly 90 in the rate of circularization due to tides raised on the star and a decrease by a factor of about 7 in the rate of circularization due to tides raised on the planet. Therefore, the present-day orbital evolution of TOI-172 b is dominated by the tides raised on its host star by the planet.

4.2 Atmospheric Characterization Prospects

The high eccentricity observed in the planet’s orbit combined with the slight evolution of the host star make TOI-172 b an interesting target for detailed characterization. While it is possible that hot Jupiters form in situ (Batygin et al. 2016), most formation theories suggest that these planets form at larger distances from their host stars (core accretion or gravitational instability; Pollack et al. 1996; Lissauer & Stevenson 2007; Boss 2000; Gammie 2001; Boley 2009) and migrate inward via two main interactions, either with drag due to the original protoplanetary disk during formation, or by gravitational interaction with another body in the system (Rasio & Ford 1996; Papaloizou et al. 2007; Fabrycky & Tremaine 2007). It was originally believed that the large number of hot Jupiters shown to have misaligned orbits relative to the host star’s spin axis indicated that these systems must migrate through gravitational scattering (Winn et al. 2010). However, the origin of these misalignments could have occurred from misalignments in the protoplanetary disk (Batygin 2012; Crida & Batygin 2014). Therefore, it is unclear what migration mechanism is responsible for close-in Jovian planets.

The migration mechanism may be revealed by studying the chemical abundances in the planet’s atmosphere. Specifically, it is more difficult to explain low carbon and oxygen abundances relative to the planet’s host star via disk migration than via disk-free migration (Madhusudhan et al. 2014). The high eccentricity of TOI-172 b is suggestive of disk-free migration, although our investigation of its orbital evolution suggests it never possessed the extremely high eccentricity that would be required to migrate from a formation location beyond the ice line. Moreover, as discussed previously, it is plausible that many hot Jupiters migrated first in the disk and then consequently through planet-planet interactions; even eccentric planets may have disk migration in their history. Nonetheless, there are currently only about a dozen planets larger than Neptune for which the eccentricity is greater than 0.2 with at least 99% confidence, a relatively short period (¡20 days), and are bright enough ($V<12$) to be well-suited for atmospheric characterization888https://exoplanetarchive.ipac.caltech.edu/ ; Akeson et al. (2013). Interestingly, less than half of these systems (including TOI-172) have a massive planetary companion ($>$3 $\,M_{\rm J}$). Therefore, TOI-172 b—with the other few known planets in this sub-sample, such as HAT-P-2b (Bakos et al. 2007) and WASP-162 (Hellier et al. 2019)—provides a great opportunity to carry out this test. If a depletion of oxygen and carbon are detected, it could provide evidence that it migrated via a disk-free method, or otherwise place constraints on its disk migration history. Future observations could try to characterize the composition of TOI-172 b’s atmosphere using current facilities like the Hubble Space Telescope and future facilities like the James Webb Space Telescope. Additionally, understanding the full architecture of the TOI-172 system, by looking for long period giant planet companions through radial velocity monitoring, may provide additional insight into its evolutionary history.

5 Conclusion

We present the discovery of TOI-172 b, a massive Jupiter in a highly eccentric $\sim$9.5 day orbit around a slightly evolved G-star. The planet has a very high density (MP = $5.42^{+0.22}_{-0.20}$ $\,M_{\rm J}$, RP = $0.965^{+0.032}_{-0.029}$ $\,R_{\rm J}$, $\rho_{P}$ = $7.53^{+0.65}_{-0.72}$ g cm*-3*) while its host star appears to be a sub-giant (M*⋆* = $1.128^{+0.065}_{-0.061}$ $\,M_{\sun}$, R*⋆* = $1.777^{+0.047}_{-0.044}$ $\,R_{\sun}$, $\log g$= $3.993^{+0.027}_{-0.028}$). Interestingly, TOI-172 b is in a rare class of highly eccentric (¿0.3), short-period ($<$20 days) massive ($>$3 $\,M_{\rm J}$) planets. The large mass and semi-major axis of TOI-172 b corresponds to a circularization timescale much larger than the age of the universe. The large eccentricity of the planet’s orbit suggests that at least some of its migration history included dynamical interactions with other components in the system. From studying the orbital evolutionary history of TOI-172   we are unable to place any useful constraints on QP or Q*⋆* since the tidal evolution is expected to be slow in this system for all reasonable values of Q*⋆* and QP. Future observations could provide more evidence for the migration mechanism by studying the atmospheric composition of TOI-172 b or by studying the entire known ensemble of hot Jupiters in the literature.

We thank Laura Kreidburg and Laura Mayorga for their valuable conversations. J.E.R. was supported by the Harvard Future Faculty Leaders Postdoctoral fellowship. AV’s contribution to this work was performed under contract with the California Institute of Technology (Caltech)/Jet Propulsion Laboratory (JPL) funded by NASA through the Sagan Fellowship Program executed by the NASA Exoplanet Science Institute. CZ is supported by a Dunlap Fellowship at the Dunlap Institute for Astronomy & Astrophysics, funded through an endowment established by the Dunlap family and the University of Toronto. CXH, JAB, and MNG acknowledge support from MIT’s Kavli Institute as Torres postdoctoral fellows. DD acknowledges support for this work provided by NASA through Hubble Fellowship grant HST-HF2-51372.001-A awarded by the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., for NASA, under contract NAS5-26555. TD acknowledges support from MIT’s Kavli Institute as a Kavli postdoctoral fellow. R.B. acknowledges support from FONDECYT Post-doctoral Fellowship Project 3180246, and from the Millennium Institute of Astrophysics (MAS). A.J. acknowledges support from FONDECYT project 1171208, CONICYT project BASAL AFB-170002, and by the Ministry for the Economy, Development, and Tourism’s Programa Iniciativa Científica Milenio through grant IC 120009, awarded to the Millennium Institute of Astrophysics (MAS). This research has made use of SAO/NASA’s Astrophysics Data System Bibliographic Services. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This work makes use of observations from the LCO network Funding for the TESS mission is provided by NASA’s Science Mission directorate. We acknowledge the use of public TESS Alert data from pipelines at the TESS Science Office and at the TESS Science Processing Operations Center. This research has made use of the NASA Exoplanet Archive and the Exoplanet Follow-up Observation Program website, which are operated by the California Institute of Technology, under contract with the National Aeronautics and Space Administration under the Exoplanet Exploration Program. This paper includes data collected by the TESS mission, which are publicly available from the Mikulski Archive for Space Telescopes (MAST). This paper includes observations obtained under Gemini program GN-2018B-LP-101. Resources supporting this work were provided by the NASA High-End Computing (HEC) Program through the NASA Advanced Supercomputing (NAS) Division at Ames Research Center for the production of the SPOC data products.