KELT-16b: A Highly Irradiated, Ultra-short Period Hot Jupiter Nearing Tidal Disruption

The KELT is an all-sky photometric survey for planets transiting bright hosts. It was originally optimized to target stars of brightnesses $8\lt V\lt 10$—filling a niche between the faintness limit of most RV surveys and the saturation limit of most transit surveys—but, counting the current discovery, it has discovered planets around stars as faint as $V\sim 12$. The survey uses two telescopes, KELT-North in Sonoita, Arizona, and KELT-South in Sutherland, South Africa. Each telescope has a 26° × 26° field of view and a 23'' pixel scale. Together these twin telescopes observe over 70% of the entire sky with 10–20 minute cadence and ∼1% photometric noise (for $V\lesssim 12;$ Pepper et al. 2007, 2012). The survey has discovered and published 13 transiting planets around bright stars (Beatty et al. 2012; Siverd et al. 2012; Pepper et al. 2013, 2016; Collins et al. 2014; Bieryla et al. 2015; Fulton et al. 2015; Eastman et al. 2016; Kuhn et al. 2016; Rodriguez et al. 2016; Zhou et al. 2016, and the current work).

KELT-16 is located in KELT-North survey field 12, centered on ($\alpha ={21}^{{\rm{h}}}{22}^{{\rm{m}}}52\buildrel{\rm{s}}\over{.} 8$, $\delta =+31^\circ 39^{\prime} 56\buildrel{\prime\prime}\over{.} 2;$ J2000). We monitored this field from 2007 to 2013 June, collecting a total of 5626 observations. One of the candidates in field 12 that passed our selection cuts was matched to TYC 2688-1839-1, located at ($\alpha ={20}^{{\rm{h}}}{57}^{{\rm{m}}}04\buildrel{\rm{s}}\over{.} 435,\delta =+31^\circ 39^{\prime} 39\buildrel{\prime\prime}\over{.} 57;$J2000). In the KELT-North light curve of this candidate, a significant box-least-squares (BLS; Kovács et al. 2002) signal was found at a period of $P\approx 0.9690039$ day with a transit depth of $\delta \approx 6.0$ millimagnitudes (mmag) and duration of approximately 2.26 hr (our image reduction and light curve processing is described in detail in Siverd et al. 2012). This prompted follow-up observations of the target (Section 2.2) and its eventual designation as KELT-16. The discovery light curve is shown in Figure 1. Properties of the host star are listed in Table 1.

Zoom In Zoom Out Reset image size

Table 1.  Stellar Properties of KELT-16

Parameter	Description	Value	References
Names	Tycho ID	TYC 2688-1839-1
	2MASS ID	2MASS J20570443+3139397
${\alpha }_{{\rm{J}}2000}$	Right Ascension (R.A.)	20:57:04.435	1
${\delta }_{{\rm{J}}2000}$	Declination (decl.)	+31:39:39.57	1
NUV	GALEX NUV mag	16.14 ± 0.26	2
${B}_{{\rm{T}}}$	Tycho ${B}_{{\rm{T}}}$ mag	12.28 ± 0.15	1
${V}_{{\rm{T}}}$	Tycho ${V}_{{\rm{T}}}$ mag	11.72 ± 0.12	1
B	APASS Johnson B mag	12.247 ± 0.062	3
V	APASS Johnson V mag	11.898 ± 0.030	3
$g^{\prime}$	APASS Sloan $g^{\prime}$ mag	11.873 ± 0.103	3
$r^{\prime}$	APASS Sloan $r^{\prime}$ mag	11.801 ± 0.030	3
$i^{\prime}$	APASS Sloan $i^{\prime}$ mag	11.685 ± 0.030	3
J	2MASS J mag	10.928 ± 0.023	4, 5
H	2MASS H mag	10.692 ± 0.022	4, 5
${K}_{{\rm{S}}}$	2MASS ${K}_{{\rm{S}}}$ mag	10.642 ± 0.016	4, 5
WISE1	WISE1 mag	10.568 ± 0.023	6, 7
WISE2	WISE2 mag	10.564 ± 0.020	6, 7
WISE3	WISE3 mag	10.352 ± 0.053	6, 7
WISE4	WISE4 mag	$\leqslant 8.787$	6, 7
${\mu }_{\alpha }$	NOMAD proper motion	5.1 ± 0.7	8
	in RA (mas yr−1)
${\mu }_{\delta }$	NOMAD proper motion	0.4 ± 0.7	8
	in DEC (mas yr−1)
RV	Absolute radial	−29.7 ± 0.1	This work
	velocity (km s−1)
$v\sin {i}_{\star }$	Stellar rotational	7.6 ± 0.5	This work
	velocity (km s−1)
Spec. Type	Spectral Type	F7V	This work
Age	Age (Gyr)	3.1 ± 0.3	This work
d	Distance (pc)	399 ± 19	This work
AV	Visual extinction (mag)	0.04 ± 0.04	This work
Ua	Space motion (km s−1)	−5.4 ± 1.3	This work
V	Space motion (km s−1)	−13.5 ± 0.4	This work
W	Space motion (km s−1)	4.1 ± 1.3	This work

Note.

^aU is positive in the direction of the Galactic Center.

References. (1) Høg et al. (2000), (2) Bianchi et al. (2011), (3) Henden et al. (2015), (4) Cutri et al. (2003), (5) Skrutskie et al. (2006), (6) Wright et al. (2010), (7) Cutri et al. (2014), and (8) Zacharias et al. (2004).

Download table as:  ASCIITypeset image

We note that this KELT candidate was initially incorrectly matched to TYC 2688-1883-1, located at ($\alpha ={20}^{{\rm{h}}}{57}^{{\rm{m}}}01\buildrel{\rm{s}}\over{.} 59,\delta =+31^\circ 39^{\prime} 37\buildrel{\prime\prime}\over{.} 74$, J2000). It was not until the first follow-up observation (Section 2.2) that it was discovered that the transit event was actually in the neighboring star TYC 2688-1839-1. This erroneous identification was a casualty of the catalog matching process, which is performed independently for the east and west (pre- and post-meridian flip) KELT data (Siverd et al. 2012). For each object identified during point-spread function (PSF) photometry, we query the Tycho-2 catalog for stars within a $6^{\prime\prime}$ radius. The east KELT object lies 2292 from TYC 2688-1883-1 and 2318 from TYC 2688-1839-1, while the corresponding west KELT object lies 1818 from TYC 2688-1883-1 and 2359 from TYC 2688-1839-1. As a result, both objects matched to the closer TYC 2688-1883-1, and thus to each other. Errors like this underscore the challenges of matching catalogs generated from high-resolution data to lower-quality images, and the value of the KELT Follow-up Network (FUN) (Section 2.2).

We obtained follow-up time-series photometry of KELT-16 to verify the planet detection, check for false-positives, and better resolve the transit profile. These observations employed the KELT-North FUN, a closely-knit international collaboration of approximately 20 privately and publicly funded observatories with ∼0.25 to 1 m aperture telescopes. The member observatories span a wide range of longitudes and latitudes to maximize the temporal and sky coverage of transit events. Scientific and logistical coordination within the network is made possible by several custom software tools including the TAPIR package (Jensen 2013), which is used to predict transit events.

KELT-FUN obtained 19 full transit observations in the BVRI and g'r'i'z' filter sets between 2015 May and 2015 December from 10 different member observatories: Canela's Robotic Observatory (CROW), Kutztown University Observatory (KUO), the University of Louisville's Moore Observatory Ritchey-Chrétien (MORC) telescope, the Manner-Vanderbilt Ritchey-Chrétien (MVRC) telescope located on the Mt. Lemmon summit of Steward Observatory, Brigham Young University's Orson Pratt Observatory (Pratt), Westminster College Observatory (WCO), Wellesley College's Whitin Observatory (Whitin), and Brigham Young University's West Mountain Observatory (WMO). The transit observations are summarized in Table 2 and the light curves shown in Figures 2 and 3. Figure 4 shows all follow-up light curves combined and binned in five minute intervals. This combined and binned light curve is not used for analysis, but rather to showcase the overall statistical power of the follow-up photometry.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 2.  Photometric Follow-up Observations of the KELT-16b Transit

Observatory	Location	Aperture	Date	Transit	Filter	FOV	Pixel Scale	FWHMb	Exposure	Cyclec	Duty Cycled	rmse	PNRf
		(m)	(UT 2015)	Epocha		(' × ')	(''pixel${}^{-1}$)	('')	(s)	(s)	(%)	(10${}^{-3}$)	(10${}^{-3}\,$min${}^{-1}$)
MVRC	Ariz.	0.6	May 23	−84	$r^{\prime}$	26 × 26	0.39	5	100	112	89	1.4	2.0
KUO	Pa.	0.6	May 24	−83	V	$19.5\times 13$	0.76	4.2	120	143	84	1.8	2.7
KUO	Pa.	0.6	May 24	−83	I	$19.5\times 13$	0.76	4.2	120	143	84	1.9	3.0
KUO	Pa.	0.6	May 26	−81	B	$19.5\times 13$	0.76	4.4	165	184	90	3.1	5.4
Pratt	Utah	0.4	Jun 23	−52	R	25 × 25	0.37	3.6	90	109	82	3.7	5.0
WMO	Utah	0.9	Jun 25	−50	B	$20.5\times 20.5$	0.61	2.6	40	55	73	2.2	2.1
WMO	Utah	0.9	Jun 25	−50	V	$20.5\times 20.5$	0.61	2.6	30	45	67	2.0	1.8
WMO	Utah	0.9	Jun 25	−50	R	$20.5\times 20.5$	0.61	2.6	20	35	57	2.4	1.8
WMO	Utah	0.9	Jun 25	−50	I	$20.5\times 20.5$	0.61	2.6	20	43	47	3.0	2.6
MORC	Ky.	0.6	Jul 25	−19	$z^{\prime}$	26 × 26	0.39	5	240	259	93	1.9	4.0
CROW	Portugal	0.25	Aug 4	−9	${R}_{{\rm{C}}}$	28 × 19	1.11	5.3	200	252	79	1.8	3.7
Whitin	Mass.	0.6	Nov 2	84	$i^{\prime}$	20 × 20	0.58	4.4	80	109	73	2.5	3.4
WCO	Pa.	0.35	Nov 3	85	$r^{\prime}$	24 × 16	1.44	3.8	160	182	88	1.8	3.1
MORC	Ky.	0.6	Nov 4	86	$g^{\prime}$	26 × 26	0.39	9	100	125	80	1.4	2.1
MORC	Ky.	0.6	Nov 4	86	$i^{\prime}$	26 × 26	0.39	9	100	125	80	1.2	1.7
KUO	Pa.	0.6	Nov 4	86	V	$19.5\times 13$	0.76	3.8	120	144	83	1.5	2.4
KUO	Pa.	0.6	Nov 4	86	I	$19.5\times 13$	0.76	3.8	120	144	83	1.8	2.8
Whitin	Mass.	0.6	Dec 7	120	$r^{\prime}$	20 × 20	0.58	4.0	80	93	86	3.0	3.7
Whitin	Mass.	0.6	Dec 7	120	$i^{\prime}$	20 × 20	0.58	4.0	80	93	86	3.1	3.9

Notes.

^aThe zeroth epoch is set to be the mid-transit time, ${T}_{{\rm{C}}}$, of the global fit to the radial velocity (RV) data (see Table 4). ^bThe average full width at half maximum (FWHM) of the stellar PSF. ^cThe cycle time is the mean of exposure time plus dead time during periods of back-to-back exposures. ^dThe duty cycle is the fraction of cycle time spent exposing. ^eThe rms is the root mean square scatter of the residuals to the best-fit transit model. ^fThe photometric noise rate (PNR) is calculated as $\mathrm{rms}/\sqrt{{\rm{\Gamma }}}$, where Γ is the mean number of cycles per minute (adapted from Fulton et al. 2011).

Download table as:  ASCIITypeset image

All observatories were Network Time Protocol or GPS synchronized to sub-millisecond accuracy, and all times have been converted to barycentric Julian dates at mid-exposure, ${\mathrm{BJD}}_{\mathrm{TDB}}\$(Eastman et al. 2010). All data were processed using the AstroImageJ package (AIJ; Collins & Kielkopf 2013; Collins et al. 2016).³⁸

We used the Tillinghast Reflector Echelle Spectrograph (TRES; Szentgyorgyi & Fűrész 2007; Fűrész et al. 2008, pp. 287–290), on the 1.5 m telescope at the Fred Lawrence Whipple Observatory (FLWO) on Mt. Hopkins, Arizona, to obtain high-resolution spectra. We obtained a total of 20 spectra between UT 2015 May 29 and 2015 December 8. The spectra have an average S/N per resolution element (SNRe) of 39 at the peak of the continuum near the Mg b triplet at 519 nm. The spectra have a resolving power of $R\sim {\rm{44,000}}$ and were extracted as described by Buchhave et al. (2010).

We derive relative radial velocities (RVs) by cross-correlating each observed spectrum order by order against the strongest observed spectrum from the wavelength range 4460–6280 Å. The observation used as the template has an RV of 0 km s⁻¹ by definition. The data are reported in Table 3.

Table 3.  TRES Relative Radial Velocity Measurements

${\mathrm{BJD}}_{\mathrm{TDB}}\$	RV	RV Error	Bisector	Bisector Error
	(m s−1)	(m s−1)	(m s−1)	(m s−1)
2457171.92841	−777.4	50.5	28.9	26.0
2457203.85614	−881.5	48.5	28.6	48.1
2457323.60011	213.8	57.0	27.9	36.7
2457345.65687	−224.6	45.4	−30.5	17.8
2457346.62517	−249.1	36.2	2.5	26.2
2457347.67298	−37.1	32.8	−35.0	23.7
2457348.65228	0.0	25.8	−4.8	26.2
2457349.63889	43.8	40.0	22.1	19.1
2457350.60616	72.8	31.2	−30.1	17.0
2457351.58680	120.8	38.5	19.6	19.0
2457354.63919	49.8	42.9	−1.9	31.6
2457355.64983	−41.8	37.1	14.3	24.5
2457356.62390	−109.2	36.6	−25.3	18.0
2457357.65057	−203.0	25.8	6.6	28.3
2457358.60143	−150.5	35.0	21.6	32.0
2457360.59367	−289.3	31.7	−35.8	22.1
2457361.59428	−451.3	36.0	−8.4	20.7
2457362.64672	−610.7	50.9	5.6	13.2
2457363.63593	−460.9	86.1	12.5	56.2
2457364.61912	−737.6	40.5	−18.4	22.3

Download table as:  ASCIITypeset image

The absolute RV of the system is determined using the absolute velocity of the star in the spectrum with the highest SNRe, which is also the spectrum used as the template when deriving the relative RVs. The absolute RV is adjusted to the International Astronomical Union (IAU) Radial Velocity Standard Star system (Stefanik et al. 1999) by first taking the RV determined by cross-correlating the spectrum with the best synthetic spectrum match near the Mg b order (see Section 3.1 for more details) and then adding the gamma velocity from the orbital solution using the relative RVs. A correction of −0.61 km s⁻¹—determined from extensive observations of the IAU standard stars—is included to correct for the non-inclusion of the gravitational redshift in the synthetic spectrum. The absolute velocity of the star is found to be −29.7 km s⁻¹ ± 0.1 km s⁻¹, where the uncertainty is an estimate of the residual systematic errors in the IAU system.

We derive values for the line profile bisector spans using procedures outlined in Buchhave et al. (2010). The bisector values are reported in Table 3 and shown in Figure 6. We measure line bisector spans to check for variations from background blends (Mandushev et al. 2005) or star spots (Queloz et al. 2001). We expect bisector variations caused from these astrophysical phenomena to vary in phase with the RV variations. There is no indication that the periodic signal is due to any astrophysical phenomena other than the orbital motion.

Zoom In Zoom Out Reset image size

We obtained natural guide star adaptive optics (AO) images of KELT-16 on 2015 December 27 with the NIRC2 instrument (Matthews & Soifer 1994) on Keck II. Twelve 15 s exposures were taken in the ${K}_{{\rm{S}}}$ band using the full NIRC2 array (1024 × 1024 pixels) with the narrow camera setting ($0\buildrel{\prime\prime}\over{.} 010$ pixel⁻¹). The full width at half maximum (FWHM) of the target star PSF was ∼5 pixels = 0050. We calibrated and removed image artifacts using dome flat fields and dark frames. Figure 7 shows a stacked image of the system using the three best frames, while all 12 frames are used for the analysis. A stellar companion is detected roughly $0\buildrel{\prime\prime}\over{.} 7$ east of the target star.

Zoom In Zoom Out Reset image size

We measure the flux ratio and on-sky separation of the companion by fitting a two-peak PSF to the data. Each peak is modeled as a combination of a Moffat and Gaussian function within an aperture of radius 2× FWHM. Details on the PSF fitting routine can be found in Ngo et al. (2015). We integrate the best-fit PSF over the same aperture to calculate the flux ratio and calculate the difference between centroids to find the separation and position angle (PA) of the companion. For the position measurement, we apply the new NIRC2 astrometric corrections from Service et al. (2016) to correct for the NIRC2 array's distortion and rotation. We find the flux ratio of the primary to secondary to be 56.5 ± 5.5, or ${\rm{\Delta }}{K}_{{\rm{S}}}=4.4\pm 0.1$ magnitudes. Adding this to the 2 Micron All-Sky Survey (2MASS)-measured ${K}_{{\rm{S}}}$ of the primary (Table 1) yields a secondary brightness of ${K}_{{\rm{S}}}=15.0\pm 0.1$ mag. The companion is separated from the KELT-16 primary by $0\buildrel{\circ}\over{.} 7177\pm 0\buildrel{\circ}\over{.} 0015$ at a PA of $95\buildrel{\circ}\over{.} 16\pm 0\buildrel{\circ}\over{.} 22$.

We also compute a $5\sigma$ contrast curve. To do so, we divide the stacked image into a series of annuli of width equal to the primary star's FWHM, where each annulus is used to compute our sensitivity at a given distance from the primary star. Then, for every pixel, we compute the sum of the flux of all neighboring pixels within a FWHM × FWHM = 5 pixel × 5 pixel box. The standard deviation of these values within the same annulus is the $1\sigma$ contrast for that annulus. We divide the limiting contrast flux by the total flux in a 5 pixel × 5 pixel box centered on the primary star to get a magnitude difference. Figure 8 shows the resulting $5\sigma$ contrast curve in ${\rm{\Delta }}{K}_{{\rm{S}}}$.

Zoom In Zoom Out Reset image size

We obtain stellar parameters from the observed TRES spectra (Section 2.3) using the Stellar Parameter Classification (SPC) tool (Buchhave et al. 2012). SPC cross-correlates an observed spectrum against a grid of synthetic spectra based on Kurucz atmospheric models (Kurucz 1992, 1979). The weighted average results are: ${T}_{\mathrm{eff}}=6227\pm 55$ K, log(${g}_{\star }$) $=\,4.03\pm 0.11$, $[m/H]=-0.01\pm 0.08$, and $v\sin {i}_{\star }=7.6\pm 0.5$ km s⁻¹. These values were calculated by taking an average of the stellar parameters determined for each spectrum individually and then weighting them according to the cross-correlation function peak height.

To determine the system parameters for KELT-16 and place it in context with other known exoplanet systems, we conduct a global fit of our follow-up photometric and spectroscopic observations. The global fit uses a modified version of EXOFAST which we refer to as multi-EXOFAST (Eastman et al. 2013)³⁹ to constrain ${R}_{\star }$ and ${M}_{\star }$ using either the Yonsei–Yale (YY) stellar evolution models (Demarque et al. 2004) or the empirical Torres relations (Torres et al. 2010). The RV measurements from TRES (Section 2.3) and raw KELT-FUN follow-up light curves (with detrending parameters—see Collins et al. 2014) are used as inputs for the fit, and the $[\mathrm{Fe}/{\rm{H}}]$ and ${T}_{\mathrm{eff}}$ (with errors) determined from the SPC analysis of the TRES spectra (Section 3.1) and the period determined from the follow-up light curves are used as priors. The blended flux from KELT-16b's nearby companion as determined by spectral energy distribution (SED) modeling (Section 3.4) is subtracted from each follow-up light curve. (See Siverd et al. 2012 for a more detailed description of the global fit procedure.)

For both the YY models and the Torres relations, we run a global fit with eccentricity constrained to zero and another with eccentricity as a free parameter, for a total of four fits. The results are shown in Tables 4 and 5. All four fits are consistent with each other to 1σ; for our interpretation and discussion in this paper, we adopt the YY circular fit. The best-fitting linear ephemeris is ${T}_{{\rm{C}}}=2457247.24791\pm 0.00019$ ${\mathrm{BJD}}_{\mathrm{TDB}}\$and $P=0.9689951\pm 0.0000024$ day. Using the standards of Pecaut & Mamajek (2013), the best-fitting ${T}_{\mathrm{eff}}$ corresponds to a stellar spectral type of F7V.

Table 4.  Median Values and 68% Confidence Intervals for the Physical and Orbital Parameters of the KELT-16 System

Parameter	Units	Adopted Value	Value	Value	Value
		(YY Circular)	(YY Eccentric)	(Torres Circular)	(Torres Eccentric)
Stellar Parameters
M⋆	Mass (${M}_{\odot }$)	${1.211}_{-0.046}^{+0.043}$	${1.206}_{-0.048}^{+0.047}$	${1.232}_{-0.058}^{+0.060}$	${1.224}_{-0.059}^{+0.061}$
R⋆	Radius (${R}_{\odot }$)	${1.360}_{-0.053}^{+0.064}$	${1.329}_{-0.070}^{+0.073}$	${1.390}_{-0.062}^{+0.063}$	${1.355}_{-0.076}^{+0.077}$
L⋆	Luminosity (${L}_{\odot }$)	${2.52}_{-0.22}^{+0.27}$	${2.40}_{-0.27}^{+0.29}$	${2.62}_{-0.25}^{+0.27}$	${2.49}_{-0.29}^{+0.31}$
${\rho }_{\star }$	Density (cgs)	${0.679}_{-0.081}^{+0.079}$	${0.72}_{-0.10}^{+0.12}$	${0.647}_{-0.070}^{+0.084}$	${0.694}_{-0.094}^{+0.12}$
$\mathrm{log}{g}_{\star }$	Surface gravity (cgs)	${4.253}_{-0.036}^{+0.031}$	${4.272}_{-0.042}^{+0.043}$	${4.242}_{-0.032}^{+0.034}$	${4.262}_{-0.041}^{+0.043}$
${T}_{\mathrm{eff}}$	Effective temperature (K)	6236 ± 54	${6237}_{-55}^{+54}$	6235 ± 53	${6235}_{-54}^{+53}$
$[\mathrm{Fe}/{\rm{H}}]$	Metallicity	$-{0.002}_{-0.085}^{+0.086}$	${0.002}_{-0.085}^{+0.086}$	${0.002}_{-0.079}^{+0.078}$	${0.002}_{-0.079}^{+0.080}$
Planet Parameters
e	Eccentricity	⋯	${0.034}_{-0.025}^{+0.047}$	⋯	${0.031}_{-0.022}^{+0.045}$
${\omega }_{\star }$	Argument of periastron (degrees)	⋯	$-{65}_{-55}^{+43}$	⋯	$-{64}_{-67}^{+52}$
P	Period (days)	0.9689951 ± 0.0000024	${0.9689951}_{-0.0000024}^{+0.0000025}$	0.9689951 ± 0.0000024	0.9689951 ± 0.0000025
a	Semimajor axis (au)	${0.02044}_{-0.00026}^{+0.00024}$	${0.02041}_{-0.00028}^{+0.00026}$	0.02055 ± 0.00033	0.02051 ± 0.00033
${M}_{{\rm{P}}}$	Mass (${M}_{{\rm{J}}}$)	${2.75}_{-0.15}^{+0.16}$	${2.78}_{-0.17}^{+0.18}$	${2.79}_{-0.16}^{+0.17}$	${2.80}_{-0.18}^{+0.19}$
${R}_{{\rm{P}}}$	Radius (${R}_{{\rm{J}}}$)	${1.415}_{-0.067}^{+0.084}$	${1.384}_{-0.080}^{+0.089}$	${1.452}_{-0.079}^{+0.081}$	${1.415}_{-0.089}^{+0.093}$
${\rho }_{{\rm{P}}}$	Density (cgs)	1.20 ± 0.18	${1.30}_{-0.23}^{+0.27}$	${1.452}_{-0.079}^{+0.081}$	${1.22}_{-0.21}^{+0.27}$
$\mathrm{log}{g}_{{\rm{P}}}$	Surface gravity	${3.530}_{-0.049}^{+0.042}$	3.554 ± 0.058	3.515 ± 0.046	${3.539}_{-0.057}^{+0.059}$
${T}_{\mathrm{eq}}$	Equilibrium temperature (K)	${2453}_{-47}^{+55}$	2427 ± 64	2472 ± 53	${2443}_{-66}^{+64}$
Θ	Safronov number	0.0654 ± 0.0045	${0.0677}_{-0.0057}^{+0.0063}$	${0.0639}_{-0.0044}^{+0.0046}$	${0.0662}_{-0.0056}^{+0.0063}$
$\langle F\rangle$	Incident flux (109 erg s−1 cm−2)	${8.22}_{-0.61}^{+0.77}$	${7.86}_{-0.81}^{+0.87}$	${8.48}_{-0.70}^{+0.75}$	${8.07}_{-0.85}^{+0.89}$
RV Parameters
${T}_{{\rm{C}}}$	Time of inferior conjunction (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457247.24791 ± 0.00019	2457247.24791 ± 0.00019	2457247.24791 ± 0.00019	2457247.24791 ± 0.00019
${T}_{{\rm{P}}}$	Time of periastron (${\mathrm{BJD}}_{\mathrm{TDB}}$)	⋯	${2457246.84}_{-0.15}^{+0.12}$	⋯	${2457246.84}_{-0.18}^{+0.15}$
K	RV semi-amplitude (m s−1)	494 ± 25	501 ± 29	494 ± 25	500 ± 29
${M}_{{\rm{P}}}\sin i$	Minimum mass (${M}_{{\rm{J}}}$)	2.74 ± 0.15	${2.76}_{-0.17}^{+0.18}$	${2.77}_{-0.16}^{+0.17}$	2.79 ± 0.18
${M}_{{\rm{P}}}/{M}_{\star }$	Mass ratio	0.00217 ± 0.00011	0.00220 ± 0.00013	0.00216 ± 0.00011	0.00219 ± 0.00013
u	RM linear limb darkening	${0.6051}_{-0.0059}^{+0.0063}$	${0.6052}_{-0.0059}^{+0.0064}$	${0.6052}_{-0.0058}^{+0.0063}$	${0.6054}_{-0.0059}^{+0.0063}$
${\gamma }_{\mathrm{TRES}}$	m s−1	−353 ± 31	$-{359}_{-35}^{+34}$	−353 ± 31	−358 ± 34
$\dot{\gamma }$	RV slope (m s−1 day−1)	−0.37 ± 0.34	−0.37 ± 0.37	−0.37 ± 0.34	−0.37 ± 0.37
$e\cos {\omega }_{\star }$		⋯	${0.012}_{-0.017}^{+0.026}$	⋯	${0.010}_{-0.016}^{+0.026}$
$e\sin {\omega }_{\star }$		⋯	$-{0.024}_{-0.048}^{+0.026}$	⋯	$-{0.019}_{-0.048}^{+0.022}$

Download table as:  ASCIITypeset image

Table 5.  Median Values and 68% Confidence Intervals for the Physical and Orbital Parameters of the KELT-16 System (Continued)

Parameter	Units	Adopted Value	Value	Value	Value
		(YY Circular)	(YY Eccentric)	(Torres Circular)	(Torres Eccentric)
Primary Transit
${R}_{{\rm{P}}}/{R}_{\star }$	Radius of the planet in stellar radii	${0.1070}_{-0.0012}^{+0.0013}$	${0.1072}_{-0.0012}^{+0.0013}$	0.1074 ± 0.0013	0.1074 ± 0.0013
a/R⋆	Semimajor axis in stellar radii	${3.23}_{-0.13}^{+0.12}$	${3.30}_{-0.16}^{+0.17}$	${3.18}_{-0.12}^{+0.13}$	${3.26}_{-0.15}^{+0.17}$
i	Inclination (degrees)	${84.4}_{-2.3}^{+3.0}$	${84.4}_{-2.0}^{+2.7}$	${83.5}_{-1.9}^{+2.7}$	${83.8}_{-1.9}^{+2.6}$
b	Impact parameter	${0.32}_{-0.16}^{+0.11}$	${0.332}_{-0.15}^{+0.098}$	${0.359}_{-0.14}^{+0.087}$	${0.360}_{-0.14}^{+0.087}$
δ	Transit depth	${0.01146}_{-0.00025}^{+0.00029}$	${0.01148}_{-0.00026}^{+0.00028}$	${0.01154}_{-0.00027}^{+0.00028}$	${0.01154}_{-0.00027}^{+0.00028}$
${T}_{\mathrm{FWHM}}$	FWHM duration (days)	0.09237 ± 0.00044	0.09244 ± 0.00044	0.09248 ± 0.00044	0.09248 ± 0.00044
τ	Ingress/egress duration (days)	${0.01133}_{-0.00099}^{+0.0013}$	${0.0115}_{-0.0010}^{+0.0012}$	${0.0118}_{-0.0011}^{+0.0012}$	${0.0118}_{-0.0011}^{+0.0012}$
T14	Total duration (days)	${0.1037}_{-0.0010}^{+0.0013}$	${0.1039}_{-0.0011}^{+0.0012}$	0.1043 ± 0.0012	0.1043 ± 0.0012
${P}_{{\rm{T}}}$	A priori non-grazing transit probability	${0.2764}_{-0.0097}^{+0.012}$	${0.265}_{-0.023}^{+0.018}$	${0.281}_{-0.011}^{+0.010}$	${0.269}_{-0.023}^{+0.018}$
${P}_{{\rm{T}},{\rm{G}}}$	A priori transit probability	${0.343}_{-0.013}^{+0.015}$	${0.328}_{-0.028}^{+0.023}$	0.348 ± 0.014	${0.334}_{-0.028}^{+0.023}$
${T}_{{\rm{C}},0}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457165.85179 ± 0.00049	2457165.85182 ± 0.00049	2457165.85179 ± 0.00049	2457165.85181 ± 0.00049
${T}_{{\rm{C}},1}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457166.82114}_{-0.0010}^{+0.00097}$	2457166.8212 ± 0.0010	${2457166.82115}_{-0.0010}^{+0.00100}$	${2457166.82119}_{-0.0010}^{+0.00100}$
${T}_{{\rm{C}},2}$	Mid- transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457166.8240}_{-0.0014}^{+0.0016}$	${2457166.8240}_{-0.0013}^{+0.0016}$	${2457166.8240}_{-0.0013}^{+0.0015}$	${2457166.8241}_{-0.0013}^{+0.0015}$
${T}_{{\rm{C}},3}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457168.7572}_{-0.0019}^{+0.0016}$	${2457168.7572}_{-0.0019}^{+0.0016}$	${2457168.7571}_{-0.0019}^{+0.0016}$	${2457168.7571}_{-0.0019}^{+0.0016}$
${T}_{{\rm{C}},4}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457196.8608}_{-0.0012}^{+0.0011}$	${2457196.8608}_{-0.0012}^{+0.0011}$	${2457196.8608}_{-0.0012}^{+0.0011}$	${2457196.8608}_{-0.0012}^{+0.0011}$
${T}_{{\rm{C}},5}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457198.7984 ± 0.0011	2457198.7984 ± 0.0011	2457198.7984 ± 0.0011	2457198.7984 ± 0.0011
${T}_{{\rm{C}},6}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457198.79803}_{-0.00066}^{+0.00065}$	${2457198.79806}_{-0.00067}^{+0.00066}$	2457198.79804 ± 0.00066	${2457198.79807}_{-0.00067}^{+0.00066}$
${T}_{{\rm{C}},7}$	Mid- transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457198.79914}_{-0.00087}^{+0.00085}$	${2457198.79918}_{-0.00086}^{+0.00084}$	${2457198.79914}_{-0.00085}^{+0.00083}$	${2457198.79917}_{-0.00085}^{+0.00083}$
${T}_{{\rm{C}},8}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457198.7969}_{-0.0014}^{+0.0015}$	2457198.7969 ± 0.0013	${2457198.7969}_{-0.0012}^{+0.0013}$	${2457198.7969}_{-0.0012}^{+0.0013}$
${T}_{{\rm{C}},9}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457228.8368}_{-0.0011}^{+0.0010}$	${2457228.8368}_{-0.0011}^{+0.0010}$	${2457228.83676}_{-0.0010}^{+0.00099}$	${2457228.83676}_{-0.0010}^{+0.00100}$
${T}_{{\rm{C}},10}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457238.52801}_{-0.00075}^{+0.00077}$	${2457238.52804}_{-0.00074}^{+0.00077}$	${2457238.52803}_{-0.00074}^{+0.00076}$	${2457238.52807}_{-0.00075}^{+0.00076}$
${T}_{{\rm{C}},11}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457328.64376}_{-0.00078}^{+0.00080}$	${2457328.64380}_{-0.00078}^{+0.00081}$	${2457328.64377}_{-0.00079}^{+0.00082}$	${2457328.64381}_{-0.00080}^{+0.00082}$
${T}_{{\rm{C}},12}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457329.61152}_{-0.00066}^{+0.00067}$	2457329.61156 ± 0.00067	${2457329.61155}_{-0.00066}^{+0.00068}$	${2457329.61158}_{-0.00067}^{+0.00068}$
${T}_{{\rm{C}},13}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457330.58170}_{-0.00061}^{+0.00060}$	2457330.58173 ± 0.00061	${2457330.58172}_{-0.00061}^{+0.00060}$	${2457330.58174}_{-0.00062}^{+0.00061}$
${T}_{{\rm{C}},14}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457330.58146 ± 0.00056	${2457330.58150}_{-0.00057}^{+0.00056}$	2457330.58147 ± 0.00056	2457330.58150 ± 0.00056
${T}_{{\rm{C}},15}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457330.58154}_{-0.00069}^{+0.00067}$	${2457330.58157}_{-0.00070}^{+0.00068}$	${2457330.58153}_{-0.00069}^{+0.00067}$	${2457330.58157}_{-0.00070}^{+0.00069}$
${T}_{{\rm{C}},16}$	Mid- transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	${2457330.5826}_{-0.0010}^{+0.0011}$	${2457330.5826}_{-0.0010}^{+0.0011}$	2457330.5827 ± 0.0010	2457330.5827 ± 0.0010
${T}_{{\rm{C}},17}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457363.5265 ± 0.0011	2457363.5266 ± 0.0011	2457363.5265 ± 0.0011	2457363.5266 ± 0.0011
${T}_{{\rm{C}},18}$	Mid-transit time (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457363.5278 ± 0.0014	2457363.5278 ± 0.0014	${2457363.5278}_{-0.0014}^{+0.0013}$	${2457363.5279}_{-0.0014}^{+0.0013}$
${u}_{1{\rm{B}}}$	Linear Limb-darkening	${0.544}_{-0.012}^{+0.013}$	${0.543}_{-0.012}^{+0.014}$	${0.544}_{-0.012}^{+0.014}$	${0.544}_{-0.012}^{+0.013}$
${u}_{2{\rm{B}}}$	Quadratic Limb-darkening	${0.2356}_{-0.0090}^{+0.0078}$	${0.2359}_{-0.0091}^{+0.0078}$	${0.2354}_{-0.0090}^{+0.0078}$	${0.2355}_{-0.0090}^{+0.0079}$
${u}_{1{\rm{I}}}$	Linear Limb-darkening	${0.2231}_{-0.0055}^{+0.0064}$	${0.2235}_{-0.0056}^{+0.0065}$	${0.2231}_{-0.0056}^{+0.0065}$	${0.2235}_{-0.0056}^{+0.0065}$
${u}_{2{\rm{I}}}$	QuadraticLimb-darkening	${0.3045}_{-0.0031}^{+0.0028}$	${0.3044}_{-0.0031}^{+0.0028}$	${0.3047}_{-0.0029}^{+0.0026}$	${0.3045}_{-0.0029}^{+0.0026}$
${u}_{1{\rm{R}}}$	Linear Limb-darkening	${0.2939}_{-0.0064}^{+0.0076}$	${0.2941}_{-0.0065}^{+0.0077}$	${0.2939}_{-0.0064}^{+0.0077}$	${0.2942}_{-0.0065}^{+0.0077}$
${u}_{2{\rm{R}}}$	Quadratic Limb-darkening	${0.3148}_{-0.0030}^{+0.0026}$	${0.3148}_{-0.0030}^{+0.0026}$	${0.3149}_{-0.0029}^{+0.0025}$	${0.3148}_{-0.0029}^{+0.0025}$
${u}_{1\mathrm{Sloan}{\rm{g}}^{\prime} }$	Linear Limb-darkening	${0.4743}_{-0.0100}^{+0.011}$	${0.4742}_{-0.0100}^{+0.012}$	${0.4746}_{-0.0099}^{+0.011}$	${0.475}_{-0.010}^{+0.011}$
${u}_{2\mathrm{Sloan}{\rm{g}}^{\prime} }$	Quadratic Limb-darkening	${0.2706}_{-0.0064}^{+0.0051}$	${0.2708}_{-0.0065}^{+0.0051}$	${0.2704}_{-0.0064}^{+0.0051}$	${0.2705}_{-0.0064}^{+0.0052}$
${u}_{1\mathrm{Sloan}{\rm{i}}^{\prime} }$	LinearLimb-darkening	${0.2412}_{-0.0057}^{+0.0066}$	${0.2415}_{-0.0057}^{+0.0067}$	${0.2411}_{-0.0057}^{+0.0067}$	${0.2416}_{-0.0058}^{+0.0067}$
${u}_{2\mathrm{Sloan}{\rm{i}}^{\prime} }$	Quadratic Limb-darkening	0.3064 ± 0.0028	${0.3063}_{-0.0029}^{+0.0028}$	${0.3066}_{-0.0027}^{+0.0026}$	${0.3064}_{-0.0028}^{+0.0026}$
${u}_{1\mathrm{Sloan}{\rm{r}}^{\prime} }$	Linear Limb-darkening	${0.3145}_{-0.0067}^{+0.0079}$	${0.3147}_{-0.0068}^{+0.0080}$	${0.3146}_{-0.0067}^{+0.0080}$	${0.3148}_{-0.0068}^{+0.0080}$
${u}_{2\mathrm{Sloan}{\rm{r}}^{\prime} }$	Quadratic Limb-darkening	${0.3161}_{-0.0031}^{+0.0025}$	${0.3161}_{-0.0031}^{+0.0024}$	${0.3162}_{-0.0030}^{+0.0024}$	${0.3161}_{-0.0030}^{+0.0024}$
${u}_{1\mathrm{Sloan}{\rm{z}}^{\prime} }$	Linear Limb-darkening	${0.1899}_{-0.0051}^{+0.0059}$	${0.1901}_{-0.0052}^{+0.0060}$	${0.1898}_{-0.0051}^{+0.0059}$	${0.1902}_{-0.0052}^{+0.0059}$
${u}_{2\mathrm{Sloan}{\rm{z}}^{\prime} }$	Quadratic Limb-darkening	${0.2979}_{-0.0034}^{+0.0026}$	${0.2978}_{-0.0034}^{+0.0026}$	${0.2981}_{-0.0031}^{+0.0025}$	${0.2979}_{-0.0032}^{+0.0025}$
${u}_{1{\rm{V}}}$	Linear Limb-darkening	${0.3787}_{-0.0076}^{+0.0090}$	${0.3788}_{-0.0077}^{+0.0091}$	${0.3788}_{-0.0076}^{+0.0090}$	${0.3790}_{-0.0077}^{+0.0091}$
${u}_{2{\rm{V}}}$	Quadratic Limb-darkening	${0.3033}_{-0.0039}^{+0.0027}$	${0.3034}_{-0.0039}^{+0.0027}$	${0.3033}_{-0.0039}^{+0.0027}$	${0.3033}_{-0.0039}^{+0.0027}$
Secondary Eclipse
${T}_{{\rm{S}}}$	Time of eclipse (${\mathrm{BJD}}_{\mathrm{TDB}}$)	2457246.76341 ± 0.00019	${2457246.771}_{-0.010}^{+0.016}$	2457246.76341 ± 0.00019	${2457246.7694}_{-0.0098}^{+0.016}$
${b}_{{\rm{S}}}$	Impact parameter	⋯	${0.309}_{-0.14}^{+0.096}$	⋯	${0.339}_{-0.13}^{+0.088}$
${T}_{{\rm{S}},\mathrm{FWHM}}$	FWHM duration (days)	⋯	${0.0887}_{-0.0072}^{+0.0041}$	⋯	${0.0896}_{-0.0071}^{+0.0035}$
${\tau }_{{\rm{S}}}$	Ingress/egress duration (days)	⋯	${0.0107}_{-0.0012}^{+0.0014}$	⋯	${0.0111}_{-0.0013}^{+0.0014}$
${T}_{{\rm{S}},14}$	Total duration (days)	⋯	${0.0996}_{-0.0084}^{+0.0051}$	⋯	${0.1009}_{-0.0083}^{+0.0044}$
${P}_{{\rm{S}}}$	A priori non-grazing eclipse probability	⋯	${0.279}_{-0.010}^{+0.011}$	⋯	0.282 ± 0.011
${P}_{{\rm{S}},{\rm{G}}}$	A priori eclipse probability	⋯	${0.346}_{-0.014}^{+0.015}$	⋯	${0.350}_{-0.015}^{+0.014}$

Note. The mid-transit times ${T}_{{\rm{C}},0}$ through ${T}_{{\rm{C}},18}$ correspond to those of the 19 KELT-FUN light curves in the same order as they are listed in Table 2.

Download table as:  ASCIITypeset images: 1 2

We estimate the age of KELT-16 by fitting the YY stellar evolution model to the stellar ${T}_{\mathrm{eff}}$, log(${g}_{\star }$), $[\mathrm{Fe}/{\rm{H}}]$, and ${M}_{\star }$ determined from the circular YY case of our global fit (Section 3.2 and Table 4). The fitting process can be understood graphically as finding the intersection point on an HR diagram of the evolutionary track that best fits M and $[\mathrm{Fe}/{\rm{H}}]$ with the associated isochrone that best fits ${T}_{\mathrm{eff}}$ and log(g). The result is an age of 3.1 ± 0.3 Gyr, indicating that KELT-16 is undergoing core hydrogen fusion and is slightly past the midpoint of its lifespan on the main sequence (Figure 9). The uncertainty in age reflects only the propagation of the uncertainties in ${T}_{\mathrm{eff}}$, log(${g}_{\star }$), $[\mathrm{Fe}/{\rm{H}}]$, and ${M}_{\star }$ from the global fit, and does not include systematic or calibration uncertainties of the YY model itself.

Zoom In Zoom Out Reset image size

We estimate the distance and reddening to KELT-16 by fitting the Kurucz (1979, 1992) stellar atmosphere models to the SED. We use broad-band photometry data from the literature spanning the GALEX NUV band at 0.227 μm to the WISE 3 band at 11.6 μm (Table 1).

However, the detection of a stellar companion by AO imaging (Section 2.4) indicates that the photometry in all of these passbands is blended. The AO imaging was single-banded, measuring only the ${K}_{{\rm{S}}}$-band flux ratio of the two stars. But under the assumption that the two stars are bound, we can iteratively use the distance and reddening from a preliminary SED fit (uncorrected for blending) along with the best-fit isochronal age (Section 3.3) to determine a ${T}_{\mathrm{eff}}$ for the secondary star based on the model of Baraffe et al. (2015). This ${T}_{\mathrm{eff}}$ is then used to extrapolate the ${K}_{{\rm{S}}}$-band flux ratio to the other photometric bands in order to separate the flux contributions from the two stars (see Section 3.5).

Figure 10 shows the final best-fit stellar atmospheres for both the KELT-16 primary and secondary stars. For these fits, we fixed the ${T}_{\mathrm{eff}}$ of the primary as determined directly from spectroscopic observations (Section 3.1) and the ${T}_{\mathrm{eff}}$ of the secondary as determined from the Baraffe model and leave distance, d, and reddening, A_V, as free parameters. We find best-fit values of $d=399\pm 19$ pc and ${A}_{V}=0.04\pm 0.04$ magnitudes, where the uncertainties reflect only the propagation of the uncertainties of the measured fluxes and do not include systematic or calibration uncertainties of the stellar atmosphere models themselves.

Zoom In Zoom Out Reset image size

A stellar companion was detected near KELT-16 by high-contrast AO imaging (Section 2.4). First, we compute the likelihood that the two stars are gravitationally bound rather than a chance alignment of foreground and background objects.

Figure 11 shows the absolute and cumulative stellar number density observed by 2MASS (Skrutskie et al. 2006) as a function of ${K}_{{\rm{S}}}$ magnitude in a 1 deg² circular region surrounding KELT-16. We include both the "unconstrained" case in which all 2MASS catalog entries are counted and the case in which counts are constrained by quality flags.⁴⁰ Regardless of constraints, the ${K}_{{\rm{S}}}=15.0\pm 0.1$ companion is found to be safely brighter than the 2MASS faintness limit of ${K}_{{\rm{S}}}\sim 15.5$ for this region of sky. Within this limit, 2MASS star counts have a very high level of completeness, and can thus be used directly to calculate a model-independent alignment probability.

Zoom In Zoom Out Reset image size

In particular, we calculate the probability that the AO observations would detect by chance a star which is at least as near to the primary as the detected companion and at least as bright as the detected companion (restricting also by the measured PA of the detected companion is not necessary since the AO imaging covered all PAs). To do so, we determine the fractional area of the sky occupied by $0\buildrel{\prime\prime}\over{.} 72$ radii circles surrounding all 2MASS sources with brightnesses of ${K}_{{\rm{S}}}\leqslant 15.0$. Technically, this ${K}_{{\rm{S}}}$ bound should be taken as a function of separation to ensure it remains above the $5\sigma$ contrast curve (Figure 8), but neglecting this only makes our calculation more conservative, since it will cause the probability of chance alignment at lesser separations to be overestimated. This approach assumes that the 2MASS catalog is complete to the faintness limit, that catalog entries have at least a $1\buildrel{\prime\prime}\over{.} 44$ separation, and that there is no blending in the 2MASS reported fluxes. Although none of these is strictly true, they are very good approximations: 2MASS completeness in the ${K}_{{\rm{S}}}$-band is reported as 99.56% over the entire sky and likely to be higher in the region immediately surrounding KELT-16 due to its lack of bright targets or tiling gaps and modest angular separation from the Galactic plane ($b\sim -9^\circ$); 2MASS spatial resolution was seeing-limited and is reported to average ∼4''; and the quality flags for confusion and photometric fit help to discount blended targets (Skrutskie et al. 2006).

Even in the most conservative case, unconstrained by (i.e., ignoring) quality flags, we count 12,487 2MASS sources of ${K}_{{\rm{S}}}\leqslant 15.0$. The probability of a chance alignment of one of these stars within $0\buildrel{\prime\prime}\over{.} 72$ of the target is 0.1569%, resulting in a 99.8431% or $\sim 3.2\sigma$ confidence that the two stars are bound.

If the two stars are bound, we can compute a number of model-dependent stellar and orbital properties of the secondary. We start with the directly observed ${K}_{{\rm{S}}}$ flux ratio and angular separation and assume the same isochrone-modeled age and SED-modeled distance and extinction as the primary star (Sections 3.3 and 3.4). From these values we calculate the extinction-corrected absolute ${K}_{{\rm{S}}}$ magnitude which, when coupled with the age, can be fit to the models of Baraffe et al. (2015) to obtain ${T}_{\mathrm{eff}}$, L, R, M, and log(g). This ${T}_{\mathrm{eff}}$ can then be matched to a spectral type and colors—and thus apparent magnitudes across multiple bands—using the spectral typing standards of Pecaut & Mamajek (2013). In short, we find the companion to be a V = 19.6 M3V type red dwarf. The full stellar properties of the secondary are summarized in Table 6.

Table 6.  Properties of the KELT-16 Stellar Companion

Parameter	Description	Value	Source
Directly Observed Parameters
${K}_{{\rm{S}},1}/{K}_{{\rm{S}},2}$	Flux ratio	56.5 ± 5.5	AO obs.
${\rm{\Delta }}r$	Angular separation (mas)	717.7 ± 1.5	AO obs.
PA	Position angle (degrees)	95.16 ± 0.22	AO obs.
Modeled Stellar Parameters
B	Johnson magnitude (mag)	21.2 ± 0.1	PM13
V	Johnson magnitude (mag)	19.6 ± 0.1	PM13
J	2MASS magnitude (mag)	15.9 ± 0.1	PM13
H	2MASS magnitude (mag)	15.3 ± 0.1	PM13
${K}_{{\rm{S}}}$	2MASS magnitude (mag)	15.0 ± 0.1	AO and
			2MASS obs.
Age	Age (Gyr)	3.1 ± 0.3	Isochrone fit
d	Distance (pc)	399 ± 19	SED fit
AV	Extinction coeff. (mag)	0.04 ± 0.04	SED fit
${M}_{{K}_{{\rm{S}}}}$	Abs. 2MASS mag (mag)	7.0 ± 0.2	AO obs. and
			SED fit
${T}_{\mathrm{eff}}$	Effective temperature (K)	3420 ± 70	BHAC15
${L}_{\star }$	Luminosity (${L}_{\odot }$)	0.0107 ± 0.0037	BHAC15
${R}_{\star }$	Radius (${R}_{\odot }$)	0.295 ± 0.040	BHAC15
${M}_{\star }$	Mass (${M}_{\odot }$)	0.300 ± 0.050	BHAC15
log(g)	Surface gravity (cgs)	4.97 ± 0.05	BHAC15
$[\mathrm{Fe}/{\rm{H}}]$	Metallicity	−0.002 ± 0.086	Global fit
Spectral Type	Spectral type	M3V	PM13
Modeled Orbital Parameters
${a}_{\min ,e=0}$	Minimum circular	286 ± 14	AO obs. and
	semimajor axis (au)		SED fit
${P}_{\min ,e=0}$	Minimum circular	3940 ± 280	Global fit and
	period (year)		BHAC15
${a}_{\min ,e\to 1}$	Minimum eccentric	143 ± 7	AO obs. and
	semimajor axis (au)		SED fit
${P}_{\min ,e\to 1}$	Minimum eccentric	1390 ± 100	Global fit and
	period (year)		BHAC15

Note. All values except those directly observed by NIRC2/Keck II AO rely on the assumption that the companion is bound. All values that depend on the Baraffe et al. (2015; BHAC15) model also depend on the direct AO observations and both the isochrone and SED model fits. All values that depend on Pecaut & Mamajek (2013; PM13) standards also depend on the ${T}_{\mathrm{eff}}$ from BHAC15. The 68% confidence intervals account only for the propagation of measurement uncertainties and do not include any systematic or calibration uncertainty inherent in the models themselves.

Download table as:  ASCIITypeset image

The stars' directly observed angular separation and SED-modeled distance imply a current minimum separation of 286 ± 14 au, or higher if the primary and secondary are not in the same sky plane. From this we calculate the minimum orbital period via Kepler's third law using the primary mass as determined by the global fit (Section 3.2) and secondary mass as determined by the Baraffe et al. (2015) model. We find that the minimum orbital period could range from 3940 ± 280 year for a circular orbit to 1390 ± 100 year for a maximum eccentricity orbit (${\mathrm{lim}}_{e\to 1}$). (Note that the distribution of observed stellar binary eccentricities has been found to be independent of stellar properties and is nearly flat, although it is also possible to fit it with a broad Gaussian of ${e}_{\mathrm{avg}}$ ∼0.4; Duchêne & Kraus 2013.)

It is statistically likely that the stars are bound, but this argument could be strengthened with high-contrast imaging in multiple bands (e.g., $J-{K}_{{\rm{S}}}$) to confirm whether or not the two stars have similar photometric distances, and/or high-contrast imaging in a future epoch to confirm whether or not the two stars have common parallax and proper motion. A search of major catalogs turns up no observations of the companion in past epochs (not surprising given its faintness and small separation), and additional observations could not be obtained by the time of publication (but are being actively pursued). Figure 12 shows a model of the two stars' separation and PA with NIRC2/Keck II AO observation uncertainties, and reveals that the earliest opportunity to check for a separation difference at the $\sim 3\sigma$ level will be in mid-2018. The PA difference will not exceed $\sim 2\sigma$ in the near future.

Zoom In Zoom Out Reset image size

We examine the three-dimensional space motion of KELT-16 to determine whether its kinematics match those of one of the main stellar populations of the Galaxy. The absolute RV measured from the TRES spectra of KELT-16 is −29.7 ± 0.1 km s⁻¹ (Section 2.3) and the proper motions from NOMAD are ${\mu }_{\alpha }=5.1\pm 0.7$ mas yr⁻¹ and ${\mu }_{\delta }=0.4\pm 0.7$ mas yr⁻¹ (Table 1). Adopting the SED-modeled distance (Section 3.4) and local standard of rest as defined in Coşkunoǧlu et al. (2011), these values transform to U, V, and W space motions of −5.4 ± 1.3, −13.5 ± 0.4, and 4.1 ± 1.3 km s⁻¹, respectively. According to the classifications of Bensby et al. (2003), this gives KELT-16 a 99.4% probability of belonging to the thin disk population of the Galaxy. Furthermore, these relatively low peculiar velocities are consistent with the best-fit isochronal age of 3.1 ± 0.3 Gyr (Section 3.3).

Variations in the observed periodicity of a planet's transit, or transit timing variations (TTVs), can be caused by the gravitational perturbations of other planets in the same system, and thus serve as a means for discovering those planets (Agol et al. 2005; Holman & Murray 2005). Models predict that TTVs will usually be largest for planets in orbital mean motion resonances (MMRs; e.g., Agol et al. 2005). However, until very recently, TTVs had not been detected for hot Jupiters. This could be because hot Jupiters' co-planets are generally of low mass or in orbits unlikely to cause MMRs, or because hot Jupiters eject most of their co-planets during their inward migration (e.g., Mustill et al. 2015). However, Becker et al. (2015) recently detected a TTV of amplitude 38 s in WASP-47b driven by a Neptune-mass planet in a nearby superior orbit (and possibly also to a lesser degree by a super-Earth in an inferior orbit). This demonstrates that multi-planet systems can induce TTVs in hot Jupiters, and that these TTV amplitudes can be large enough to detect from the ground.

However, for KELT-16b, the uncertainties in the mid-transit times, ${T}_{{\rm{C}}}$, of the global fit to the individual KELT-FUN light curves range from $42\leqslant {\sigma }_{{T}_{{\rm{C}}}}\leqslant 150\,{\rm{s}}$, with only about half of them having values of ∼1 min or less (Tables 5 or 7). Even taking only this best half, KELT-16b would need to have a TTV signal of ∼180 s to be detected at the $3\sigma$ level by KELT-FUN. Such a high TTV signal is expected to be rare for a hot Jupiter, but is theoretically possible even for a low-mass perturbing planet for special combinations of that planet's eccentricity and inclination (Payne et al. 2010).

We thus check KELT-16b for TTV signals. To do so, we compute the O–C residuals between the global fit mid-transit times for the 19 KELT-FUN light curves, ${T}_{{\rm{C}}}$, and the global fit ephemeris, ${T}_{{\rm{C}}}=2457247.24791\pm 0.00019$ ${\mathrm{BJD}}_{\mathrm{TDB}}\$and $P=0.9689951\pm 0.0000024$ day (Table 4). Note that, during the global fit, the ${T}_{{\rm{C}}}$ for each light curve was fit individually and independently of the others—i.e., P was fit to the ${T}_{{\rm{C}}}$'s and not the other way around—so any TTVs that may exist would not have been altered by the global fit itself.

The results are listed in Table 7 and plotted in Figure 13. No data points are found to have a statistically significant deviation from the global fit ephemeris (the flat dashed line): the average O–C is 60 s or $0.7\sigma$, with the largest single outlier being 230 s or $1.8\sigma$. We thus do not believe these outliers are astrophysical in origin and attribute them to observational systematics as described by Carter & Winn (2009). We carefully ensured that all follow-up observations were correctly converted to BJD${}_{\mathrm{TBD}}$ (Eastman et al. 2010) and that each observatory clock was synchronized with a standard clock to sub-millisecond accuracy (Section 2.2) such that the limiting uncertainty for each ${T}_{{\rm{C}}}$ is the global fit (and therefore the quality of the photometric data) itself, rather than any systematic differences in timekeeping by observers.

Zoom In Zoom Out Reset image size

Table 7.  Transit Times for KELT-16b

Epoch	${T}_{{\rm{C}}}$	${\sigma }_{{T}_{{\rm{C}}}}$	O–C	O–C	Telescope
	(${\mathrm{BJD}}_{\mathrm{TDB}}$)	(s)	(s)	(${\sigma }_{{T}_{{\rm{C}}}}$)
−84	2457165.85179	42	−45	−1.1	MVRC
−83	2457166.82114	85	−15	−0.2	KUO
−83	2457166.8240	130	230	1.8	KUO
−81	2457168.7572	150	−180	−1.2	KUO
−52	2457196.8608	100	50	0.5	Pratt
−50	2457198.7984	90	20	0.2	WMO
−50	2457198.79803	56	−11	−0.2	WMO
−50	2457198.79914	74	85	1.1	WMO
−50	2457198.7969	120	−110	−0.9	WMO
−19	2457228.8368	90	−20	−0.2	MORC
−9	2457238.52801	65	91	1.4	CROW
84	2457328.64376	68	21	0.3	Whitin
85	2457329.61152	57	−85	−1.5	WCO
86	2457330.58170	52	17	0.3	MORC
86	2457330.58146	48	−3	−0.1	MORC
86	2457330.58154	58	3	0.1	KUO
86	2457330.5826	90	90	1.0	KUO
120	2457363.5265	90	−70	−0.8	Whitin
120	2457363.5278	120	40	0.3	Whitin

Download table as:  ASCIITypeset image

Following Penev et al. (2014), we model the orbital evolution of KELT-16b due to the dissipation of the tides raised by the planet on the the host star under the assumption of a constant phase lag. The starting configuration of the system was tuned to reproduce the presently observed system parameters (Table 4) at the assumed system age of 3.1 Gyr (Section 3.3). The evolution includes the effects of the changing stellar radius and luminosity following the YY circular stellar model with mass and metallicity as given in Table 4, but neglects the effects of the stellar rotation, assuming that the star is always rotating sub-synchronously relative to the orbit.

Orbital and stellar irradiation evolutions are shown in Figures 14 and 15 for a range of stellar tidal quality factors (${Q}_{* }^{\prime }={10}^{5},{10}^{6},{10}^{7},{10}^{8},\mathrm{and}\,{10}^{9}$), where ${Q}_{* }^{\prime -1}$ is the product of the tidal phase lag and the Love number. We find that the insolation received by the planet is well above the empirical inflation irradiation threshold ($\approx 2\times {10}^{8}$ erg s⁻¹ cm⁻²; Demory & Seager 2011) for the entire main-sequence existence of the star, except in the very early stages of stellar evolution for the case ${Q}_{* }^{\prime }={10}^{5}$ (Figure 15).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We consider a wide range of ${Q}_{* }^{\prime }$ because of the wide range of proposed mechanisms for tidal dissipation in current theoretical models and the conflicting observational constraints backing those models, especially for stars that may have surface convective zones (see the review by Ogilvie 2014 and references therein). Furthermore, because the dependence on stellar mass and tidal frequency is different for the different proposed mechanisms, we make the simplifying assumption that ${Q}_{* }^{\prime }$ remains constant over the life of the star. However, with multi-year baselines, it may be possible in the future to empirically constrain the lower limit on ${Q}_{* }^{\prime }$ for KELT-16 via precise measurements of the orbital period time decay (see Hoyer et al. 2016).

Finally, note that this model does not account in any way for the larger-distance Type II or scattering-induced migration that KELT-16b and other hot Jupiters likely undergo. It considers only the close-in migration due to tidal friction alone.

KELT-16b is in a select group of ultra-short period inflated hot Jupiters transiting relatively bright host stars. There are only five others with < 1 day periods: WASP-18b, -19b, -43b, -103b, and HATS-18b (Hellier et al. 2009; Hebb et al. 2010; Hellier et al. 2011; Gillon et al. 2014; and Penev et al. 2016, respectively), and only one of these, WASP-18b, is more massive than KELT-16b (Figures 16 and 17). Such planets are subject to extreme irradiation and strong tidal forces, making it plausible that long-term follow-up of these systems could reveal evidence of atmospheric evolution or orbital decay (e.g., Hoyer et al. 2016). Furthermore, comparative studies of such planets could inform planet formation and evolution scenarios, in particular how significant a role planet mass plays in the rate of atmospheric or dynamic evolution.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Because of its high irradiation and large size, KELT-16b has a higher infrared flux than most planets. It is thus expected to have larger phase curve amplitudes and secondary eclipse depths—and thus higher S/N—in the HST and Spitzer infrared passbands. This makes KELT-16b a promising target for tackling a number of current open questions in the field regarding exoplanetary atmospheres, such as the presence and cause of temperature–pressure inversions, the identification of upper atmosphere absorbers at hot temperatures, and day-to-night heat transfer.

Due to its high mass, we find KELT-16b to have a relatively small atmospheric scale height of $H={{kT}}_{\mathrm{eq}}/\mu {g}_{{\rm{P}}}=267\pm 15$ km, where we assume a hydrogen-dominated atmosphere with a mean molecular weight of $\mu =2$ following Winn (2010, p. 55), ${T}_{\mathrm{eq}}$ and ${g}_{{\rm{P}}}$ are taken from our global fit (Table 4), and k is the Boltzmann constant. Thus, measuring the wavelength-dependent radius during transit to produce a transmission spectrum is not as compelling with current telescopes, but KELT-16b will be a good candidate for such observations with the larger aperture JWST. Therefore, a comprehensive study of KELT-16b in both transit and eclipse will be possible in the near future.

Of particular interest, KELT-16b is a higher-mass, higher-temperature analog of the well-studied planet WASP-43b, providing an opportunity for several key differential measurements. For example, due to its high level of irradiation, KELT-16b falls into the "pM class" of planets that are warm enough to have opacity due to TiO and VO gasses, while WASP-43b may belong to the cooler "pL class" of planets in which the TiO and VO have condensed and rained out (Fortney et al. 2008). pM planets such as KELT-16b are expected to absorb incident stellar flux in the stratosphere, leading to temperature–pressure inversions and large day-to-night temperature swings (hotter days and cooler nights). pL planets, on the other hand, are expected to absorb incident flux deeper in the atmosphere where atmospheric dynamics will more readily redistribute absorbed energy, leading to less drastic day-to-night temperature swings (cooler days and hotter nights). If KELT-16b's temperature cools enough in going from the dayside to the nightside, TiO and VO could rain out at the terminator and its temperature–pressure inversion disappear on the night side (e.g., Burrows & Sharp 1999; Lee et al. 2015).

Finally, KELT-16b presents an opportunity to study the trend observed in solar system giants of decreasing metallicity with increasing planet mass. This relationship was extended to higher mass by WASP-43b (Kreidberg et al. 2014) and could be tested out to nearly 3 ${M}_{{\rm{J}}}$ with KELT-16b (Figure 18).

Zoom In Zoom Out Reset image size

Because most formation theories require that giant planets form beyond the snow line, KELT-16b may have once orbited at a distance of $\gtrsim 5\,\mathrm{au}$ with a period of ∼10 years or more (Garaud & Lin 2007; Kennedy & Kenyon 2008) before migrating inward due to Type II or scattering-driven migration (Section 1). The likely existence of a bound, widely separated stellar companion in the KELT-16 system (Section 3.5) presents the possibility that the KL mechanism influenced this migration (Kozai 1962; Lidov 1962). If KELT-16b formed beyond the snow line, its KL timescale could have been as short as ${T}_{\mathrm{KL}}\sim 2.5\,\mathrm{Myr}$ (e.g., Kiseleva et al. 1998), assuming the current minimum circular orbital period of the stellar companion (Table 6). If the stellar companion orbit is eccentric, then the minimum KL timescale could have been far less than this. Future observations of the Rossiter–McLaughlin (RM) effect could help provide evidence for past KL librations by measuring system misalignment. It is also noteworthy that KELT-16 sits nearly exactly at the ${T}_{\mathrm{eff}}=6250$ K boundary for misalignment suggested by Winn et al. (2010).

Once KELT-16b had migrated to within a few tens of stellar radii, its orbital dynamics became dominated by tidal forces. Based on the modeling of these forces, KELT-16b began a runaway in-spiral by the age of ∼1 Gyr (Section 3.8). It is currently orbiting at a radius of $a=3.373{R}_{\star }=4.382{R}_{\odot }$, just ∼1.7 times the Roche limit of ${a}_{{\rm{R}}}=2.009\pm 0.045{R}_{\star }$ $=3.373\pm 0.022{R}_{\odot }$. This is within the 2 ${a}_{{\rm{R}}}$ radius at which planets scattered inward and having eccentric orbits are expected to tidally circularize (Matsumura et al. 2010), consistent with the circular orbital solutions of the global fit (Section 3.2).

At such a close distance to its star, it is worth exploring whether KELT-16b may be in danger of either evaporation due to extreme irradiation or disintegration due to extreme tidal forces. KELT-16b's current X-ray and extreme UV flux (${F}_{\mathrm{XUV}}$)-driven atmospheric mass loss rate is expected to be $\dot{M}\lt 4.0\times {10}^{-13}\,{M}_{{\rm{J}}}$ yr⁻¹, as calculated via Equations (6) and (7) of Salz et al. (2015) using the KELT-16 system parameters derived in this work and assuming that KELT-16 has the same upper limit on ${F}_{\mathrm{XUV}}$ as measured by Salz et al. (2015) for the similar star WASP-18. Even if this rate persisted for the life of KELT-16b (untrue since the irradiation evolves; Figure 15), the total mass loss over the life of KELT-16b to date would be $\sim 0.0012\,{M}_{{\rm{J}}}$, or ∼0.045% of the planet's total mass. Thus, although KELT-16b is highly irradiated, its high overall mass limits the evaporative mass loss rate and keeps it safely above the evaporation zone as defined by the models of Kurokawa & Nakamoto (2014; approximately depicted by the shaded region in Figure 16).

Nevertheless, KELT-16b will soon be tidally shredded as it continues a runaway in-spiral toward its star. Based on the orbital evolution model of Section 3.8, KELT-16b could cross the Roche limit in a mere ∼550,000 yr if the stellar tidal quality factor is ${Q}_{* }^{\prime }={10}^{5}$, or in as long as ∼2.5 Gyr from now—roughly doubling the planet's current age and lasting nearly as long as its star—if ${Q}_{* }^{\prime }={10}^{9}$.

KELT-16b's ultra-short period offers a number of practical advantages for future observations. First, it means that observing a full orbital phase requires less telescope time—an important consideration for oversubscribed telescopes such as the HST, Spitzer, the soon-to-be-launched JWST, and others. The short period also makes it easier for most observatories to cover a higher fraction of the full phase curve in a single continuous observation. And when more telescope time is available, a short period allows for the more rapid observation and phase-folded stacking of multiple orbits to achieve a high S/N. The other P < 1 day giant planets have been heavily observed in part for these reasons. However KELT-16b has the additional advantage of its period being very close to (just 45 minutes shy of) a sidereal day, making it unusually easy to also collect multiple consecutive transit or secondary eclipse observations from the ground when it is "in phase" for a particular observatory.

In addition to its short period, KELT-16b has the eighth highest incident flux, and thus equilibrium temperature (${T}_{\mathrm{eq}}\approx 2400$ K), among giant planets transiting bright stars, giving it a large phase curve amplitude and secondary eclipse signal. Finally, its relatively bright (V = 11.7) host offers moderately high photometric precision.

KELT-16b thus has all of the characteristics to become one of the top observational targets among exoplanets.