© 2021 The Author(s) Published by Oxford University Press on behalf of Royal Astronomical Society
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 1
Planet Hunters TESS III: two transiting planets around
the bright G dwarf HD 152843
N. L. Eisner,1? B. A. Nicholson,1,2 O. Barraga´n,1 S. Aigrain,1 C. Lintott,1 L. Kaye,1
B. Klein,1 G. Miller,1 J. Taylor,1 N. Zicher,1 L. A. Buchhave,3 D. A. Caldwell,4
J. Horner,2 J. Llama,5 A. Mortier,6,7 V. M. Rajpaul,6 K. Stassun,8 A. Sporer,9
A. Tkachenko,10 J. M. Jenkins,11 D. Latham,12 G. Ricker,9 S. Seager,9,13,14 J. Winn,15
S. Alhassan,16 E. M. L. Baeten,16 S. J. Bean,16 D. M. Bundy,16 V. Efremov,16
R. Ferstenou,16 B. L. Goodwin,16 M. Hof,16 T. Hoffman,16 A. Hubert,16 L. Lau,16
S. Lee,16 D. Maetschke,16 K. Peltsch16,17 C. Rubio-Alfaro16 and G. M. Wilson16
1Department of Physics, University of Oxford, Keble Road, Oxford OX1 3RH, UK
2Centre for Astrophysics, University of Southern Queensland, Toowoomba, Queensland 4350, Australia
3DTU Space, National Space Institute, Technical University of Denmark, Elektrovej 328, DK-2800 Kgs. Lyngby, Denmark
4 SETI Institute 189 Bernardo Ave, Suite 200 Mountain View, CA 94043, USA
5 Lowell Observatory, 1400 W. Mars Hill Rd., Flagstaff, AZ 86001, USA
6Astrophysics Group, Cavendish Laboratory, University of Cambridge, J.J. Thomson Avenue, Cambridge CB3 0HE, UK
7Kavli Institute for Cosmology, University of Cambridge, Madingley Road, Cambridge CB3 0HA, UK
8Department of Physics and Astronomy, Vanderbilt University, Nashville, TN 37235, USA
9Department of Physics and Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology,
Cambridge, MA 02139, USA
10 Institute of Astronomy, KU Leuven, Celestijnenlaan 200D, 3001 Leuven, Belgium
11 NASA Ames Research Center, Moffett Field, CA 94035, USA
12 Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138, USA
13Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
14Department of Aeronautics and Astronautics, MIT, 77 Massachusetts Avenue, Cambridge, MA 02139, USA
15 Department of Astrophysical Sciences, Princeton University, 4 Ivy Lane, Princeton, NJ 08544, USA
16 Citizen Scientist, Zooniverse c/o University of Oxford, Keble Road, Oxford OX1 3RH, UK
17 School of Computer Science & Technology, Algoma University, Sault Ste. Marie, Ontario, P6A 2G4, Canada
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We report on the discovery and validation of a two-planet system around a bright (V = 8.85 mag)
early G dwarf (1.43 R , 1.15 M , TOI 2319) using data from NASAaˆA˘Z´s Transiting Exoplanet
Survey Satellite (TESS ). Three transit events from two planets were detected by citizen scientists
in the month-long TESS light curve (sector 25), as part of the Planet Hunters TESS project.
Modelling of the transits yields an orbital period of 11.6264+0.0022−0.0025 days and radius of 3.41
+0.14
−0.12 R⊕ for
the inner planet, and a period in the range 19.26–35 days and a radius of 5.83+0.14−0.14 R⊕ for the outer
planet, which was only seen to transit once. Each signal was independently statistically validated,
taking into consideration the TESS light curve as well as the ground-based spectroscopic follow-up
observations. Radial velocities from HARPS-N and EXPRES yield a tentative detection of planet b,
whose mass we estimate to be 11.56+6.58−6.14 M⊕, and allow us to place an upper limit of 27.5 M⊕
(99% confidence) on the mass of planet c. Due to the brightness of the host star and the strong
likelihood of an extended H/He atmosphere on both planets, this system offers excellent prospects
for atmospheric characterisation and comparative planetology.
Key words: methods: statistical - planets and satellites: detection - stars: funda-
mental parameters - stars:individual (TIC 349488688, HD 152843)
? E-mail: nora.eisner@new.ox.ac.uk
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
2 Eisner et al.
1 INTRODUCTION
Systems with multiple transiting planets offer a wealth of
information for exoplanetary science. In particular they
allow for comparative planetology: studying planets that
have formed out of the same material, but have formed
and evolved in different environments, receiving different
amounts of incident flux from the host star, resulting in
differing masses, radii and composition. Well characterised
multi-planet systems therefore provide important model
constraints that single-planet systems cannot, providing
insight into planetary system architecture and evolution-
ary pathways, as well as informing ongoing planet popula-
tion studies (e.g, Tremaine & Dong 2012; Dietrich & Apai
2020).
The Kepler mission (Borucki et al. 2010) revealed
that multi-planetary systems are common (Latham et al.
2011), with almost half of all Kepler planets listed in the
NASA Exoplanet Archive belonging to multi-planet systems
(Akeson et al. 2013). However, the majority of the hundreds
of multi-planet systems found by Kepler are too faint to
follow-up with ground-based high-resolution spectroscopy.
This has resulted in most known multi-planet systems lack-
ing well determined masses, densities, bulk compositions and
atmospheric characterisation, all of which are key to helping
us understand the overall planet population.
NASA’s Transiting Exoplanet Survey Satellite (TESS;
Ricker et al. 2015), however, targets stars that are on av-
erage a 30-100 times brighter than those observed by the
Kepler mission, thus allowing us to follow up and constrain
the properties of systems that were previously inaccessi-
ble. TESS has already discovered tens of previously un-
known, multi-planet systems (e.g., Gandolfi et al. 2019a;
Quinn et al. 2019; Dragomir et al. 2019a; Gilbert et al.
2020; Mann et al. 2020; Fridlund et al. 2020; Carleo et al.
2020; Leleu et al. 2021).
Detecting transiting multi-planet systems with longer-
period planets is challenging due to the reduced transit
probability of those planets, as well as the challenges asso-
ciated with detecting planets showing single transits using
automated detection algorithms. For this reason, alternative
methods are often used to identify longer-period, single tran-
sit candidates, such as machine learning (e.g., Pearson et al.
2018; Zucker & Giryes 2018), or visual vetting with the help
of citizen science (Eisner et al. 2020a; Fischer et al. 2012).
Furthermore, verifying the planetary nature of single
transit objects is challenging, as the lack of a known orbital
period complicates follow-up efforts. However, this is made
easier in the situation of multi-planet systems. Latham et al.
(2011) and Lissauer et al. (2012) independently showed that
systems with multiple planet candidates are statistically
less likely to be false positives, compared to single planet
systems. This is helpful to consider in following up single-
transit, longer-period planets with closer companions which
are themselves more easily verifiable as true planetary com-
panions.
Despite the large number of exoplanet discoveries made
by TESS and Kepler, systems with more than one tran-
siting planet around stars brighter than V ∼ 10 (the typ-
ical magnitude required for atmospheric follow-up, e.g.,
Fortenbach & Dressing 2020) containing planets with mea-
sured masses remain exceedingly rare. As of April 2021,
there are only 17 transiting planets (in 12 systems) with
mass measurements better than 50% precision around stars
with V < 10 listed in the NASA Exoplanet Archive
(Akeson et al. 2013). A list of these systems and their cor-
responding parameters can be found in Appendix A. Sig-
nificant observing resources have been, and continue to be,
devoted to each of them.
In this paper we present a new multi-planet sys-
tem, with the discovery of two planets orbiting around
HD 152843. These candidates were initially identified in
TESS Sector 25 by citizen scientists taking part in the
Planet Hunters TESS project (Eisner et al. 2020a). In Sec-
tion 2 we outline the discovery of the candidates and the
vetting tests carried out based on the TESS photometric
light curve. In Section 3 we discuss the spectroscopic data
obtained with HARPS-N and EXPRES and in Section 4 we
discuss the joint photometric and spectroscopic data anal-
ysis. Finally, the results are discussed in Section 5 and the
conclusions presented in Section 6.
2 TESS PHOTOMETRY
HD 152843 was observed by TESS only in Sector 25 of the
primary mission. The spacecraft obtained images at a ca-
dence of two-seconds, which were combined on board into
two-minute cadence data products. These were processed
and reduced by the Science Processing Operations Center
(SPOC; Jenkins et al. 2016). Throughout this work we use
the pre-search data conditioning (PDC) light curve from the
SPOC pipeline, as shown in Figure 1. The data gap seen in
the centre of the full light curve corresponds to the time
taken (∼ 1 day) for the spacecraft to send the data to Earth
and re-orient itself. The black dashed lines at the bottom
of the figure indicate the times of the periodic momentum
dumps caused by the firing of the thrusters as the spacecraft
adjusts the spin rate of the reaction wheels approximately
every 5.5 days.
2.1 Discovery of HD 152843 b and HD 152843 c
The light curve shown in Figure 1 exhibits three
transit events belonging to different transiting planets,
with HD 152843 b shown in blue and HD 152843 c
shown in pink. The first transit event of HD 152843 b
(TBJD−2457000 ∼1994.28 d) and the single transit event of
HD 152843 c (TBJD−2457000 ∼2002.77 d) were flagged as a sin-
gle Threshold Crossing Event (TCE) by the SPOC pipeline,
as two events caused by the same ‘object’. However, due
to the different depths of these two transits the TCE was
not promoted to TESS Object of Interest (TOI) status, due
to the assumption that the two events correspond to the
primary and secondary eclipses of an eclipsing binary. The
second transit event of HD 152843 b was not flagged by the
pipeline.
All three transit events were identified by the Planet
Hunters TESS (PHT) citizen science project (Eisner et al.
2020a). PHT, which is hosted by the Zooniverse platform
(Lintott et al. 2008, 2011), harnesses the power of over 25
thousand registered citizen scientists who visually vet all
of the TESS two-minute cadence light curves in search for
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 3
transit events that were ignored or missed by the main tran-
sit detection pipeline and other teams of professional as-
tronomers. The light curve of HD 152843 was seen by 15 cit-
izen scientists, 12 of whom identified all three transit events,
and 3 who identified only two out of the three events. The
target was initially brought to the attention of the PHT re-
search team via the PHT discussion forum 1. We uploaded
both planet candidates to the Exoplanet Follow-up Observ-
ing Program for TESS (ExoFOP-TESS) site on 2020-08-07
as a community TESS Object of Interest (cTOI). The in-
ner planet has since been promoted to the priority 1 (1 =
highest priority, 5 = lowest priority) candidate TOI 2319.01.
2.2 Excluding false positive scenarios
Astrophysical and instrumental false positives are common
in the TESS data, in particular due to the large (21 ”/pix)
pixel scale. We used the publicly available Lightcurve Anal-
ysis Tool for Transiting Exoplanets (latte; Eisner et al.
2020b) in order to perform standard diagnostic tests that
help to rule out false positive scenarios including background
eclipsing binaries, systematic effects, and background events
such as asteroids passing through the field of view. For a
full description of the diagnostic tests we refer the reader to
Eisner et al. (2020b), however in brief the tests include:
(i) Checking that the transit events do not coincide with
the times of the periodic momentum dumps.
(ii) Checking that the x and y centroid positions are
smoothly varying with time in the vicinity of the transit
events.
(iii) Examining light curves of the 5 nearest two-minute
cadence TESS stars to check for systematic effects.
(iv) Examining light curves extracted for each pixel sur-
rounding the target in order to ensure that the signal is not
the result of a background eclipsing binary, a background
event or caused by systematics.
(v) Checking that there are no spurious signals, such as
sudden jumps or strong variations, in the background flux.
(vi) Comparing transit shapes and depths when extracted
with different aperture sizes.
(vii) Comparing between the average in-transit and av-
erage out-of-transit flux, as well as the difference between
them.
(viii) Checking the location of nearby stars brighter than
V-band magnitude 15 as queried from the Gaia Data Release
2 catalog (Gaia et al. 2018).
(ix) Performing the box-Least-Squares fit to search for
additional signals.
Tests (i) to (iv) enabled us to rule out events caused
by systematic effects due to the satellite or instrument, and
tests (iii) to (viii) increased our confidence that the signals
are not caused by astrophysical false positives, such as blends
where the photometric aperture of a bright target contains
a faint eclipsing binary.
As blends are common in the TESS data, we
searched for nearby Gaia Data Release 2 catalog stars
(Gaia Collaboration 2018) within 110 arcseconds of the tar-
get, and found there to only be a single star with a V-band
1 https://www.zooniverse.org/projects/nora-dot-eisner/planet-hunters-tess/talk/2112/1552434?comment=2520798
magnitude brighter than 15, as shown by the orange circle
in Figure 2, where the red star shows HD 152843 and the
red outline highlights the aperture used to extract the light
curve.
In order to rule out this nearby star as the cause
of the transit events, we calculated the magnitude differ-
ence between HD 152843 and the faintest companion star
that could plausibly be responsible for the observed transit
shapes and depths. Following the methodology outlined by
Vanderburg et al. (2019) and the transit parameters derived
using pyaneti (see Section 4.4) we show that the maximum
magnitude difference between the target star and a possible
background contaminant is 1.5 magnitude in the V band.
This allows us to confidently conclude that the 14.4 magni-
tude star (5.6 magnitude fainter than HD 152843), located
at an angular separation of ∼ 31.3 ”, is not responsible for
either of the planetary signals.
2.3 Limits on additional planets
We quantify the detectability of additional planets in the
TESS light curve using a transit injection and recovery test
(e.g., Eisner et al. 2020c). In brief, we removed the known
transit events prior to injecting synthetic signals into the
PDC TESS light curve. The injected signals were generated
using the batman package (Kreidberg 2015), with planet
radii ranging from 1 to 12 R⊕ and periods ranging from 3
to 24 days, both sampled at random from a log-uniform dis-
tribution. The impact parameter and eccentricity were as-
sumed to be zero throughout and we used a quadratic limb-
darkening law with q1 and q2 of 0.16 and 0.59, respectively,
as taken from Table 15 in Claret (2016) using the stellar
parameters given in Table 1. Once the signals were injected,
we used an iterative non-linear filter (Aigrain & Irwin 2004)
to estimate and subtract residual systematics on timescales
> 1.7 days.
We simulated and injected a total of 750,000 transit
events. The Box Least Squares (BLS; Kova´cs et al. 2002)
algorithm was then used to try to recover the injected sig-
nals. The BLS search sampled a frequency grid that was
evenly-spaced from 0.01 to 1 day−1. For each simulation, we
recorded the period and orbital phase corresponding to the
highest peak in the BLS periodogram. If the recovered or-
bital period and phase agreed to within 1 % of the injected
period, the signal was deemed to be correctly identified. The
completeness, assessed in radius and period bins with width
of 0.25 R⊕ and 0.75 d respectively, was then taken to be the
fraction of correctly identified transit signals.
The results, shown in Figure 3, highlight, as expected,
that the automated BLS search is strongly biased towards
detecting shorter period planets that transit multiple times
in the light curve. The limited duration of the TESS obser-
vations of ∼ 27 d, interrupted by a 1.3 d data gap, results
in a sharp decline in completeness for periods longer than
around 13 days. For planets greater than 2 R⊕ we recover
94 per cent of signals with periods between 12 and 13 days
and 78 per cent of signals with periods between 14 and 15
days. The completeness for the parameters of planet b is
close to 100 %, while the completeness for the parameters of
planet c is close to 0% due to the fact that there is only one
transit within the available TESS light curve. We caution
that the simulated signals were injected into the PDC light
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
4 Eisner et al.
Figure 1. Flux time series for HD 152843 vs TESS Julian day (BJD-2457000.0) for Sectors 25. The light grey points show the short
cadence data with a 2 minute sampling, whilst the black points are 10 minute averages. The dashed vertical lines at the bottom of the
figure show the times of the TESS momentum dumps. The transit events are shown in blue and pink, corresponding to the inner and
outer planet candidates.
Table 1. Stellar parameters.
Parameter Value Source
Identifiers
HD 152843
TOI 2319
TIC 349488688 Stassun et al. (2019)
Gaia DR2 4564566554995619072 Gaia eDR3(a)
2MASS J16550834+2029287 2MASS(b)
Astrometry
αJ2000 16:55:08.373 Gaia eDR3
(a)
δJ2000 20:29:29.509 Gaia eDR3
(a)
Distance (pc) 107.898 ± 0.317 Bailer-Jones et al. (2018)
pi (mas) 9.161 ± 0.015 Gaia eDR3(a)
Photometry
B 9.380 ± 0.020 Tycho-2 (c)
V 8.850 ± 0.010 Tycho-2 (c)
J 7.896 ± 0.018 2MASS(b)
H 7.655 ± 0.016 2MASS(b)
K 7.629 ± 0.020 2MASS(b)
W1 7.563 ± 0.031 WISE(d)
W2 7.594 ± 0.020 WISE(d)
W3 7.607 ± 0.019 WISE(d)
Physical Properties
Spectral Type G0
Effective Temperature Teff (K) 6310 ± 100 This work
Surface gravity log g? (cgs) 4.19 ± 0.03 This work
v sin i?( km s
−1) 8.38 ± 0.50 This work
[M/H] (dex) −0.22 ± 0.08 This work
[Fe/H] (dex) −0.16 ± 0.05 This work
vmic ( km s−1) 1.66 ± 0.13 This work
vmac ( km s−1) 2 Bruntt et al. (2010)
Stellar mass M? (M) 1.15 ± 0.04 This work
Stellar radius R? (R) 1.43 ± 0.02 This work
Stellar density ρ? (ρ) 0.40 ± 0.03 This work
Star age (Gyr) 3.97 ± 0.75 This work
Note – (a) Gaia early Data Release 3 (eDR3; Gaia Collaboration et al. 2020). (b) Two-micron All Sky Survey (2MASS; Cutri et al.
2003). (c) Tycho-2 catalog (Høg et al. 2000). (d) Wide-field Infrared Survey Explorer catalog (WISE; Cutri & et al. 2013)
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 5
Figure 2. The median TESS image around HD 152843. The
aperture used to extract the light curve is shown by the red outline
and the orange dot depicts the location of the only star brighter
than V = 15 within 110 arcseconds of the target (red star), as
queried by Gaia DR2 (Gaia et al. 2018). This nearby star (V =
14.4) is located at an angular separation of ∼ 31.3 ”.
Figure 3. The recovery completeness of injected transit signals
into the light curve of HD 152843 as a function of the radius and
orbital period. The signals were recovered using a BLS search.
The properties of HD 152843 b and HD 152843 c are shown by
the red and yellow star respectively.
curve, which has already undergone detrending and system-
atics corrections by the SPOC pipeline. The presented re-
covery rates are, therefore, systematically higher than one
might otherwise expect if the signals had been injected into
the raw light curve (e.g., Lienhard et al. 2020). Overall,
this analysis highlights the difficulties associated with de-
tecting longer-period planets using automated algorithms,
and demonstrates a need for alternative detection methods
such as citizen science.
3 SPECTROSCOPIC DATA
3.1 Reconnaissance spectra
We made use of the Las Cumbres Observatory (LCO) tele-
scopes with the Network of Robotic Echelle Spectrographs
(NRES, Brown et al. 2013). This fibre-fed spectrograph,
mounted on a 1.0-m telescope, has a resolution of R = 53,000
and a wavelength coverage of 380 to 860 nm. We obtained
two spectra of HD 152843 on the 15th and 22nd August 2020
with per pixel signal to noise ratios (SNR) of 38 and 25 at
520 nm, respectively. The two spectra gave radial velocity
estimates of 9.7 ± 0.2 km/s and 9.6 ± 0.7 km/s, which are
consistent within their uncertainties, and thus allowed us to
rule out the possibility that the transit events are caused by
an eclipsing binary.
3.2 High-resolution spectra
We acquired high-resolution (R≈ 115 000) spectra with the
High Accuracy Radial velocity Planet Searcher in the North-
ern hemisphere (HARPS-N; Cosentino et al. 2012, 2014)
spectrograph mounted at the 3.6-m Telescopio Nazionale
Galileo in La Palma, Spain, via Director’s Discretionary
Time (program ID A41DDT4). We obtained 18 spectra be-
tween 5 September and 11 November 2020 (mean SNR ∼ 89
at at 550 nm). Each spectrum has simultaneous wavelength
calibration with a Fabry-Perot etalon and was reduced
via the standard HARPS Data Reduction Software (DRS;
Baranne et al. 1996) using a G2 spectral template (mean
RV uncertainty ∼ 4.2 m s−1). Additionally, we extracted the
HARPS-N RV measurements using the TERRA pipeline
(Anglada-Escude´ & Butler 2012), which uses a template-
matching approach based on a template generated by stack-
ing all of the spectra. The results extracted using DRS
and TERRA have comparable uncertainties, with a slightly
larger root-mean-square scatter in the TERRA extracted
data. Around 71% of the DRS/TERRA RVs agree within
1σ and around 82% agree within 2σ. For the remainder of
our analysis we used the data extracted with the DRS.
We derived the log R′HK values for the HARPS-N spectra
with SNR> 100 using the calibrations of Noyes et al. (1984),
and found the values to range from -4.96 to -4.94 with a
mean value of -4.95. This low value suggests that HD 152843
is a quiet star. We also note that there is no correlation
between the log R′HK values and the radial velocities.
In addition to the HARPS-N observations we obtained
22 spectra between 9 September and 10 October 2020 using
the high-resolution (R≈ 150 000) EXtreme PREcision Spec-
trometer (EXPRES; Jurgenson et al. 2016; Petersburg et al.
2020; Blackman et al. 2020) mounted on the 4.3-m Lowell
Discovery Telescope (LDT; Levine et al. 2012), USA. Each
spectrum was calibrated using a Thorium Argon lamp and
a stabilized Laser Frequency Comb and the RVs were ex-
tracted using the EXPRES analysis pipeline (for detail see
Petersburg et al. 2020). Due to poor seeing and high air-
mass, 12 of those spectra (with SNR < 25 at 550 nm) were
not used for further analysis. The mean SNR and mean RV
uncertainty of the used spectra are ∼ 82 and ∼ 9.5 m s−1,
respectively. All HARPS-N and EXPRES RV measurements
are listed in Table 2.
4 DATA ANALYSIS
4.1 Stellar atmospheric parameters
The fundamental stellar parameters of HD 152843, namely
the effective temperature (Teff), surface gravity (log g),
metallicity ([M/H]), projected rotational velocity (vsin i),
and microturbulent velocity (ξt), were extracted using three
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
6 Eisner et al.
Table 2. Radial velocity measurements.
Time RV σRV SNR Source
(BJD-2457000) ( m s−1) ( m s−1)
2098.3521 4.2460 1.7400 155.1 HARPS-N
2101.6407 -3.4170 12.1000 19.0 EXPRES*
2101.6553 19.1100 13.8290 16.0 EXPRES*
2101.6701 -27.3050 14.6120 14.0 EXPRES*
2101.6849 -26.5790 14.0830 14.0 EXPRES*
2102.3412 -4.4874 2.3570 117.5 HARPS-N
2102.6207 11.1370 11.9490 20.0 EXPRES*
2102.6351 10.9350 10.7270 21.0 EXPRES*
2102.6519 15.8470 12.4080 20.0 EXPRES*
2102.6656 -0.8630 11.5030 23.0 EXPRES*
2102.6843 26.9160 11.7640 22.0 EXPRES*
2102.6999 -24.1910 11.2430 22.0 EXPRES*
2102.7140 -15.1840 12.8060 18.0 EXPRES*
2102.7312 -43.8920 14.5010 13.0 EXPRES*
2104.3651 -8.9263 18.0372 19.9 HARPS-N*
2110.3253 -3.5140 3.2192 86.6 HARPS-N
2111.3788 1.8856 2.7648 99.8 HARPS-N
2117.3242 5.3618 2.9210 95.2 HARPS-N
2119.3255 5.5608 3.2501 78.5 HARPS-N
2120.3307 0.2539 2.2039 126.2 HARPS-N
2120.4134 3.4885 3.3055 85.9 HARPS-N
2120.6143 5.4070 4.9410 95.0 EXPRES
2123.5929 -0.2250 5.3170 81.0 EXPRES
2123.6069 0.1490 4.9690 83.0 EXPRES
2125.3192 -1.6649 2.4766 110.9 HARPS-N
2126.3165 -6.6282 4.4202 64.7 HARPS-N
2126.5970 -1.0650 8.9100 41.0 EXPRES
2126.6118 11.0790 6.8700 57.0 EXPRES
2127.3185 4.9884 3.9503 71.8 HARPS-N
2128.3180 9.1121 3.1712 87.2 HARPS-N
2129.5838 5.0890 5.1920 92.0 EXPRES
2129.5967 13.1530 5.8300 64.0 EXPRES
2130.3156 1.4459 6.3500 45.3 HARPS-N
2130.5850 9.9870 5.1790 90.0 EXPRES
2132.5928 8.5890 4.4480 114.0 EXPRES
2132.6078 5.3240 4.8730 110.0 EXPRES
2152.2939 1.4524 2.5161 112.1 HARPS-N
2153.2902 -4.2017 2.4502 115.7 HARPS-N
2154.2902 3.2138 3.6046 80.8 HARPS-N
2155.2908 -11.5868 7.1983 44.0 HARPS-N
Note – * indicates that the spectrum was not used for further
independent methods: ARES+MOOG 2, Grid Search in
Stellar Parameters (gssp) 3, and Stellar Parameter Clas-
sification (SPC).
The ARES+MOOG method derives stellar atmospheric
parameters using a curve-of-growth method based on the
equivalent widths (EW) of the Fe I and Fe II lines (for de-
tails see Sousa 2014). The EWs of the spectral lines were
automatically extracted from a stacked spectrum of all of
the HARPS-N data (with SNR > 45), using the Ares2
code (Sousa et al. 2015). The stacked spectrum has a SNR
∼ 350 at 6000 . The radiative transfer code MOOG (Sneden
1973) was then used to extract the stellar parameters, as-
suming local thermodynamic equilibrium (LTE) and using
2 ARESv2: http://www.astro.up.pt/~sousasag/ares/; MOOG
2017: http://www.as.utexas.edu/~chris/moog.html
3 GSSP: https://fys.kuleuven.be/ster/meetings/binary-2015/gssp-software-package
a grid of ATLAS plane-parallel model atmospheres (Kurucz
1993). The value of log g was subsequently further refined
(Mortier et al. 2014) and systematic and precision errors
combined in quadrature. The method yields the following
values: Teff = 6348 ± 100 K, log g = 4.31 ± 0.12, [Fe/H] =
-0.16 ± 0.06, and ξt = 1.82 ± 0.13 km s−1.
We also used the open access gssp code (Tkachenko
2015), which compares the normalised observed spectrum
with a grid of synthetic spectra. A stacked spectrum of all
of the HARPS-N data (with SNR > 45) was used for this
analysis. The goodness of fit of each synthetic spectrum was
assessed using a χ2 metric. The atmospheric models used as
part of this code were pre-computed using the LLmodels
software (Shulyak et al. 2004) and the code assumed LTE.
We independently optimised the abundances of Fe, Mg, Ti,
Cr and Ni. The best-fit spectral model is shown in Figure 4.
In order to d termine the best-fit parameters and abun-
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 7
Figure 4. Section of the stacked HARPS-N spectra with SNR
> 45 (black) and the best-fit model as determined and com-
puted with the GSSP software (red). The parameters and abun-
dances of this best-fit model, combined with the results from the
ARES+MOOG and SPC analysis, were used to determine the
stellar parameters listed in Table 1.
dances, the χ2 value was recorded for each combination of
parameters. The projected χ2 values were then fit with a
fourth order polynomial for each parameter in order to de-
termine the global minimum, which corresponds to the value
of the best-fit parameter. The uncertainties were taken as
the intersection between the polynomial and the 1 σ un-
certainty limit. The following atmospheric parameters were
obtained using GSSP: Teff = 6368 ± 100 K, log g = 4.16 ±
0.10, [M/H] = -0.17 ± 0.05, [Fe/H] = -0.16 ± 0.05, v sin i =
8.56 ± 0.5 km s−1and ξt = 1.50 ± 0.15 km s−1. We note that
the derived v sin i value is not representative of the true rota-
tional velocity of the star; instead, it represents a combined
line broadening due to rotation and macroturbulence. Since
we do not rely on the rotation rate of the star in our subse-
quent analysis, we find disentangling the effects of rotation
and macroturbulent velocity to be beyond the scope of this
study.
Finally, we used the SPC tool (for details see
Buchhave et al. 2012, 2014). Similarly to GSSP, SPC uses
spectral synthesis, which was independently carried out on
each HARPS-N spectrum (where SNR > 45). We obtained
the following values: Teff = 6175 ± 50 K, log g = 4.15 ± 0.10,
[M/H] = -0.26 ± 0.08, and v sin i = 8.2 ± 0.5 km s−1.
The values listed in Table 1, the averages of the results
obtained from these three methods, were used for all subse-
quent analysis. Finally, we note that the spectra show almost
no sign of Ca H and K re-emission, suggesting low magnetic
activity.
4.2 SED fitting
As an independent determination of the basic stellar pa-
rameters, we performed an analysis of the broadband spec-
tral energy distribution (SED) of the star together with
the Gaia DR2 parallax (adjusted by +0.08 mas to ac-
count for the systematic offset reported by Stassun & Torres
2018), in order to determine an empirical measurement of
the stellar radius, following the procedures described in
Stassun & Torres (2016); Stassun et al. (2017, 2018). We
pulled the BTVT magnitudes from Tycho-2, the JHKS magni-
0.1 1.0 10.0
λ (μm)
-12
-11
-10
-9
lo
g 
λF
λ 
(er
g s
-
1  
cm
-
2 )
Figure 5. Spectral energy distribution of HD 152843. Red sym-
bols represent the observed photometric measurements, where the
horizontal bars represent the effective width of the passband. Blue
symbols are the model fluxes from the best-fit Kurucz atmosphere
model (black).
tudes from 2MASS, the W1–W4 magnitudes from WISE, the
GGBPGRP magnitudes from Gaia, and the FUV and NUV
magnitudes from GALEX. Together, the available photom-
etry spans the full stellar SED over the wavelength range
0.15–22 µm (see Figure 5).
We performed a fit using Kurucz stellar atmosphere
models, with Teff , [Fe/H], and log g adopted from the spec-
troscopic analysis. The remaining free parameter is the ex-
tinction AV , which we limited to the maximum line-of-sight
value from the Galactic dust maps of Schlegel et al. (1998).
The resulting fit (Figure 5) has a reduced χ2 of 1.9; the re-
duced χ2 improves to 1.1 if we exclude the GALEX FUV
flux, which exhibits a modest UV excess suggestive of chro-
mospheric activity. We find a best-fit AV = 0.04+0.05−0.04.
Integrating the (unreddened) model SED gives the bolo-
metric flux at Earth Fbol = 7.72 ± 0.18 × 10−9 erg s−1 cm−2.
Taking the Fbol and Teff together with the Gaia paral-
lax gives the stellar radius, R? = 1.42 ± 0.05 R. In ad-
dition, we can estimate the stellar mass from the spec-
troscopic log g together with R? from above, giving M? =
1.11± 0.15 M, which is consistent with that empirical rela-
tions of Torres et al. (2010), giving M? = 1.22 ± 0.07 M.
Finally, we can use the star’s rotation and mild UV
excess (Fig. 5) to estimate an age via empirical rotation-
activity-age relations. The observed FUV excess implies a
chromospheric activity of log R′HK = −4.51± 0.05 via the em-
pirical relations of Findeisen et al. (2011), which in turn im-
plies a stellar rotation period of Prot = 5.0 ± 0.9 d via the
empirical relations of Mamajek & Hillenbrand (2008), con-
sistent with the upper limit Prot/sin i = 8.7 d obtained from
the spectroscopic v sin i and R?.
4.3 Stellar mass, radius, age, and distance
The stellar parameters were extracted using isochrones and
stellar evolutionary tracks. For this analysis, the combined
ARES+MOOG, GSSP and SPC effective temperature and
metallicity were used as inputs, along with the Gaia eDR3
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
8 Eisner et al.
parallax, and the magnitude of the star in eight bands. All
of the values used for this analysis are presented in Table 1.
For an in depth discussion of this analysis see
Mortier et al. (2020), however, in brief, this analysis made
use of the isochrones package (Morton 2015a), using stel-
lar models from the Dartmouth Stellar Evolution Database
and from the MESA isochrones and Stellar Tracks (MIST;
Choi et al. 2016). We used MultiNest (Feroz et al. 2019)
for the likelihood analysis and 400 live points. The analy-
sis was run six times: for each of the stellar models (Dart-
mouth/MIST) it was run three times using the Teff and
metallicity from the spectroscopic analysis (Section 4.1).
The stellar values were extracted from the combined pos-
teriors, taking the median and the 16th and 84th quantiles.
The stellar mass, radius, density and age are listed in Ta-
ble 1.
4.4 Joint transit and RV modelling
The transit and RV data were jointly analysed using the
open access pyaneti code (Barraga´n et al. 2019). In brief,
pyaneti creates marginalised posterior distributions for dif-
ferent parameters by sampling the parameter space using a
Markov chain Monte Carlo (MCMC) approach. We use the
limb-darkened quadratic models by Mandel & Agol (2002)
to fit the flattened transits. The RV data are fit with Kep-
lerian RV models.
We first modelled the transits. Since planet c transits
only once, the two planets were analysed independently. For
planet b both transits were fitted simultaneously. This al-
lowed us to fit for transit epoch, orbital period, impact fac-
tor, scaled planet radius, and scaled semi-major axis.
The single transit event (planet c) was modelled by fit-
ting for the same parameters as for planet b, with the ex-
ception of the orbital period and scaled semi-major axis, as
these cannot be constrained in the case of a single transit
event. Instead, we obtained a possible period range of 13 to
35 days at the 99% confidence interval, using the relations
presented in Osborn et al. (2016) and assuming a circular
orbit. These results were used to create uniform priors for
all the transit model parameters, for a joint RV and transit
analysis.
All fitted parameters and priors used for the joint mod-
eling are presented in Table 3. We note that for this analysis
we allow the orbits to be eccentric in order to give more flex-
ibility to the analysis. We sample for the stellar density ρ?,
and we recover the scaled semi-major axis for each planet
in the system using Kepler’s third law. We use a Gaussian
prior on ρ? using the stellar mass and radius derived in Sec-
tion 4.3. We also note that because planet c only exhibits a
single-transit event we use a wide uniform prior on its pe-
riod, based on the results from the single-transit analysis.
However, we truncated the lower period limit at 19.26 d, as
a shorter orbital period would have necessarily resulted in
further transit events being present within the TESS light
curve.
We sampled the parameter space using an MCMC ap-
proach with 500 independent chains and created posterior
distributions using 5000 iterations of converged chains with
a thin factor of 10. This generated a posterior distribution
made with 250,000 independent samples for each parameter.
The fitted parameters extracted from such posteriors can be
Figure 6. Corner plot for Kb , Pc , and Kc . First row in each
column shows the posterior distribution (blue line) together with
the prior shape (solid green line). Vertical solid (red) lines show
the median, and vertical dashed (red) lines indicate 68.3% credi-
ble intervals. The rest of sub-plots show the correlation between
parameters.Transparent blue points show individual samples and
solid black lines show iso-density contours.
found in Table 3. We note that the model and data only
weakly constrain the orbital period of HD 152843 c, Pc . Fur-
thermore, posterior distributions for the semi-amplitudes of
both planets, Kb and Kc , are truncated at zero. These pos-
teriors and their correlations are shown in Figure 6.
The posterior of Kb corresponds to a 2σ detection,
3.09+1.76−1.66 m s
−1, while planet c is not detected with an up-
per limit of 5.6 m s−1, at 99% confidence level. Figures 7
and 8 show the derived transit and RV models, respectively,
together with the corresponding data.
4.5 Statistical Validation
The open source python package VESPA was used to calculate
the statistical false positive probability (FPP) of both the
planet candidates (Morton 2012, 2015b; Morton et al. 2016).
In brief, VESPA computes the probabilities of a number of as-
trophysical scenarios that could result in the transit events
using a Bayesian framework. These consist of HEB (hier-
archical eclipsing binary), EB (eclipsing binary) and BEB
(background eclipsing binary). A population of stars is sim-
ulated for each scenario using the TRILEGAL galactic model
(Girardi et al. 2005) and the shape of the simulated transits
compared to the transits in the observed TESS light curve.
This results in a likelihood for each false positive scenario.
The FPPs for HD 152843 b and HD 152843 c are 0.05 %
and <0.001 %, respectively, meaning that they are both
below the traditionally required threshold of FPP < 1 %
(Morton et al. 2016; Crossfield et al. 2016). We also note
that the VESPA model does not consider multiplicity in planet
systems, which has been shown to decrease the FPP by
at least an order of magnitude (Lissauer et al. 2011, 2012,
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 9
Table 3. System parameters.
Parameter Prior(a) Value(b) Comments
Model Parameters for HD 152843b
Orbital period Porb (days) U[11.5, 11.7] 11.6264+0.0022−0.0025
Transit epoch T0 (BJD - 2457000) U[1994.25, 1994.30] 1994.2831+0.0024−0.0029
Parametrization e sinω U[−1, 1] −0.11+0.19−0.28 The code ensures e < 1
Parametrization e cosω U[−1, 1] −0.07+0.37−0.38 The code ensures e < 1
Scaled planet radius Rp/R? U[0, 0.1] 0.02201+0.00081−0.00073
Impact parameter, b U[0, 1.1] 0.32+0.27−0.20
Doppler semi-amplitude, K ( m s−1) U[0, 50] 3.09+1.76−1.66 2σ detection
Model Parameters for HD 152843c
Orbital period Porb (days) U[19.26, 35] 24.38+6.23−3.4 Truncated posterior (see Fig. 6)
Transit epoch T0 (BJD - 2457000) U[2002.73, 2002.8] 2002.7708+0.0011−0.0011
Parametrization e sinω U[−1, 1] 0.05+0.19−0.21 The code ensures e < 1
Parametrization e cosω U[−1, 1] 0.04+0.38−0.37 The code ensures e < 1
Scaled planet radius Rp/R? U[0, 0.1] 0.03764+0.00069−0.00074
Impact parameter, b U[0, 1.1] 0.49+0.10−0.11
Doppler semi-amplitude, K ( m s−1) U[0, 50] 7.1 Upper limit (99% interval of the posterior)
Other Parameters
Stellar density ρ? (g cm−3) N[0.56, 0.04] 0.568+0.042−0.043
Parameterized limb-darkening coefficient q1 U[0, 1] 0.183+0.156−0.09 q1 parameter as in Kipping (2013)
Parameterized limb-darkening coefficient q2 U[0, 1] 0.47+0.35−0.31 q2 parameter as in Kipping (2013)
Offset velocity HARPS-N ( km s−1) U[−0.50, 0.50] 0.0007+0.0013−0.0012
Offset velocity EXPRES ( km s−1) U[−0.50, 0.50] 0.006+0.0021−0.0021
Jitter HARPS-N ( m s−1) U[0, 100] 3.02+1.47−1.27
Jitter EXPRES ( m s−1) U[0, 100] 1.06+1.88−0.82
Jitter TESS (ppm) U[0, 500] 39+35−27
Derived parameters HD 152843b
Planet mass (M⊕) · · · 11.56+6.58−6.14 2σ detection
Planet radius (R⊕) · · · 3.41+0.14−0.12
Planet density ρ (g cm−3) · · · 1.58+0.96−0.83
Semi-major axis a (AU) · · · 0.1053+0.003−0.0031
Eccentricity e · · · 0.14+0.25−0.10 Upper limit of 0.72 (99% interval of the posterior)
Transit duration τ (hours) · · · 5.53+0.11−0.11
Orbit inclination i (deg) · · · 88.85+0.73−0.73
Insolation Fp (F⊕) · · · 255.7+21.6−19.7
Derived parameters HD 152843c
Planet mass (M⊕) · · · 27.5 Upper limit (99% interval of the posterior)
Planet radius (R⊕) · · · 5.83+0.14−0.14
Planet density ρ (g cm−3) · · · 0.82 Upper limit (99% interval of the posterior)
Eccentricity e · · · 0.115+0.173−0.08 Upper limit of 0.59 (99% interval of the posterior)
Transit duration τ (hours) · · · 6.359+0.087−0.071
Orbit inclination i (deg) · · · 88.89+0.18−0.15
Note – (a) U[a, b] refers to uniform priors between a and b, N[a, b] to Gaussian priors with mean a and standard deviation b. (b)
Inferred parameters and errors are defined as the median and 68.3% credible interval of the posterior distribution.
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
10 Eisner et al.
4 2 0 2 4
0.998
0.999
1.000
1.001
Re
la
tiv
e 
flu
x
Error 
 bar
TOI-2319b
TOI-2319b binned
0.002
0.000
0.002
Re
sid
ua
ls
4 2 0 2 4
0.998
0.999
1.000
1.001
Re
la
tiv
e 
flu
x
Error 
 bar
TOI-2319c
TOI-2319c binned
4 2 0 2 4
T-T0 (hours)
0.002
0.000
0.002
Re
sid
ua
ls
Figure 7. Phase-folded TESS light curve of HD 152843 b (upper
panel) and HD 152843 c (lower panel). Nominal TESS data are
shown in light gray together with 10-min binned data in solid
colour. The inferred transit model for each planet is over-plotted
with a solid black line. An example of the nominal white noise in
the data is also shown.
2014). Lissauer et al. (2012), for example, estimated that
systems with two or more planets in the Kepler data were 25
times less likely to be false positives. Furthermore, the de-
rived upper mass limits of both planets enable us to rule out
that the events are caused by an eclipsing binary. As both
planet candidates reach the required threshold of 99 % con-
fidence level we consider both HD 152843 b and HD 152843 c
statistically validated.
5 RESULTS AND DISCUSSION
The inner planet HD 152843 b (Pb =11.6264+0.0022−0.0025 d)
has a radius of Rb = 3.41+0.14−0.12 R⊕while the outer planet
HD 152843 c has a radius of Rc = 5.83+0.14−0.14 R⊕. The ra-
dial velocity measurements allowed us to constrain the mass
of the innermost planet to Mb = 11.56+6.58−6.14 M⊕ and derive
an upper mass limit of the outer planet (i.e. of planet c)
of Mc < 27.5 M⊕. Even though the obtained spectroscopic
data do not provide a 3 − σ detection of the mass of either
planet, the derived upper mass limits allow us to confirm
that the transit signals seen in the TESS light curve are not
the result of an eclipsing binary. Furthermore, they allow us
to make predictions about future photometric and spectro-
scopic follow-up observations (see Sections 5.1 and 5.2).
While the orbital period of the inner planet is well de-
termined, based on the two transit events seen in the TESS
light curve, this is not the case for the singly transiting outer
planet. We, therefore, constrain Pc based on the minimum
period allowed by the TESS light curve, the transit duration
and shape, and the joint modeling of the transit and RVs.
As shown in Figure 6, the joint modeling of the light
curve and the RVs produce a truncated posterior distri-
bution for Pc. This distribution favours orbital periods of
around 23 days. While this could indicate a 2:1 mean motion
resonance (MMR) with planet b, this could also be an arte-
fact introduced into the modeling by planet b. Furthermore,
while we can rule out orbital periods shorter than 19.26 d, it
is possible that HD 152843 c has an orbital period of, or close
to, 19.375 d, which would be a 5:3 MMR with HD 152843 b.
The dynamical stability of these orbits and the effects of
resonances in multi-planet systems is further discussion in
Sections 5.3 and 5.1, respectively.
In order to place HD 152843 into a wider context, Fig-
ure 9 shows the position of planet b and c in the radius-
insolation diagram alongside all known exoplanets (grey
points). Multi-planet systems with measured masses around
stars brighter than V = 10 are shown by the orange cir-
cles (see Appendix A for more detail on these systems).
HD 152843 b and HD 152843 c are depicted by the blue tri-
angle and pink square, respectively. The figure highlights
a noticeable lack of well characterised multi-planet systems
around bright stars, which are key for comparative atmo-
spheric studies. Furthermore, it shows that the planet c lies
in a sparsely populated region of parameter space. This
makes it valuable, as the characterisation of planets in this
underpopulated region of parameters can help constrain the-
ories of planet formation and evolution.
The two planets also stand out in terms of their bulk
densities. Given the minimum radius and upper mass limit
of HD 152843 c, this planet has a density < 0.82 g cm−3, sug-
gesting that the planet has an extended gaseous envelope.
Similarly, the density of HD 152843 b is 1.58+0.96−0.83 g cm
−3,
making both planets prime candidates for atmospheric char-
acterisation, as discussed further in Section 5.4.
One possible explanation for the expected low density of
planet c is that it formed at a greater distance from the host
star prior to migrating to its current orbit. This would have
allowed the planet to accrete a significant H/He envelope,
due to the colder and less dense gas present farther away
from the host star. Furthermore, planets that undergo this
type of migration are often found to be the outer planets in
mean-motion resonant chains (Lee & Chiang 2016). Future
spectroscopic and photometric observations will allow us to
further constrain the orbital period of planet c in order to
determine whether the two planets are in resonance with one
another.
Alternatively, the two planets could have formed in
situ and their differing planet properties resulted from
subsequent diverging evolutionary pathways. For example,
extreme ultraviolet irradiation from the host star could
have enabled atmospheric loss through photoevaporation of
the inner planet (Owen & Wu 2016; Chen & Rogers 2016),
stripping it of its extended gaseous envelope, while the outer
planet could have been inflated, resulting in the observed low
density of planet c.
Theory also suggests that the low density of the plan-
ets could be due to tidal heating, which could result in an
increase in entropy (e.g., Millholland 2019) and thus an in-
flated radius. Finally, Gao & Zhang (2020) and Wang & Dai
(2019) independently suggest that the apparent radii could
be enhanced by photochemical hazes in the atmospheres,
resulting in an underestimate of the densities of planets. Fu-
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 11
Figure 8. RV time-series (upper panel) and phase-folded RV plots for HD 152843 b (lower left panel) and HD 152843 c (lower right panel)
following the subtraction of the instrumental offsets. HD 152843 c plot has been phase folded using a period of 24.5 days. HARPS-N (red
diamonds) and EXPRES (blue circles) RV measurements along with their nominal uncertainties are shown in each panel. The vertical
grey lines mark the error bars including jitter. Solid black lines show the respective inferred model.
ture transmission spectra of planet c, for example at mid-
infrared wavelengths where the atmosphere is less affected
by hazes, will allow us to differentiate between different for-
mation scenarios and therefore provide useful constraints for
theoretical models of planet formation and migration.
5.1 Transit Timing Variations prospects
Transit Timing Variations or TTVs are often observable in
multi-planet systems as two planets dynamically interact,
as predicted by Agol et al. (2005) and Holman & Murray
(2005). This is especially the case when planets are near
orbital resonance, which is potentially true for HD 152843.
Measuring TTVs, especially when combined with RV data
allows for the refinement of planetary mass and orbital pa-
rameters, critical for interpreting atmospheric transmission
spectra in smaller planets (Batalha et al. 2019). It can also
enable the detection of inclined non-transiting planets and
can therefore lend insight into system demographics and ar-
chitectures (Brakensiek & Ragozzine 2016).
TTVs were assessed for this system using the best-fit
planetary parameters across a range of mass, period, and
eccentricity solutions using the TTVFast framework of n-
body simulations (Deck et al. 2014). Maximum likelihood
solutions for the periods of planets b and c indicate a pos-
sible 2:1 resonance, which would result in TTVs with an
amplitude ranging from 5-40 minutes, and a super period of
Figure 9. Planet insolation-radius diagram of confirmed exo-
planets from the NASA Exoplanet Archive (grey points, retrieved
April 2021). Orange points show members of systems with more
than one planet, with mass measurements better than 50 % and
around stars brighter than V = 10 (Akeson et al. 2013, see Ap-
pendix A). The black lines connect planets that are within the
same system. Planets that are not connected by a black line are
in multi-systems where only one planet has a mass measurement
with better than 50 % accuracy. HD 152843 b and HD 152843 c
are shown by the blue triangle and pink square, respectively.
approximately 2-3 years, allowing for follow-up observations
to detect discernible TTVs on the scale of about a year. This
amplitude would be greatly increased for non-zero eccen-
tricities. In the window of possible period solutions, further
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
12 Eisner et al.
resonant solutions include a 5:3 resonance; however, signif-
icant TTVs would not be observed away from resonance.
Followup studies of this system should enable us to signifi-
cantly constrain planetary masses, eccentricities, and other
orbital parameters, given both the presence or absence of
significant TTVs.
5.2 Rossiter-McLaughlin effect prospects
The moderate projected rotational velocity of HD 152843
(v sin i ∼ 8.2 km s−1) makes it a good candidate for study-
ing the Rossiter-McLaughlin effect (RM; Rossiter 1924;
McLaughlin 1924), which provides an estimate of the spin-
orbit alignment of the orbiting planets with the host star
(e.g., Schneider 2000). The RM effect helps to shed light onto
the dynamical history of the system, as mechanisms such
as planet-disk interactions help to preserve the initial spin-
orbit alignment, while planet-planet interactions promote
misalignment (e.g., Chatterjee et al. 2008; Deeg et al. 2009;
Storch et al. 2017). The number of multi-planet systems
with measured obliquities remains small (e.g., Hjorth et al.
2021; Dalal et al. 2019). We estimate the RM effect to be
3.71+0.89−0.74 m s
−1 and 9.56+2.65−2.7 m s
−1 for HD 152843 b and c,
respectively (Winn 2010). Future precision RV observations
(for example we obtained a typical precision of 4 m s−1for
this target with HARPS-N) will be able to detect the RM
of planet c, thus allowing for the determination of the true
obliquity of the target.
5.3 Orbital dynamics
Given the uncertainty around the period of planet c, we are
unable to perform a full dynamical analysis of the system,
as in the work of e.g. Horner et al. (2019). However, we can
estimate the system stability by comparing the possible pe-
riod scenarios of planet c to the general cases presented by
Agnew et al. (2019).
In general, those authors found that dynamical stability
can be broken into three broad regimes: highly stable orbits
(when the two orbits do not approach more closely than
several mutual Hill radii; and when the two orbits are more
widely spaced than the 1:2 mean-motion resonance); quali-
fied stability (when the orbits are closer together than the
1:2 resonance, but have stability ensured by mutual mean-
motion resonance) and likely strong instability (which typi-
cally occurs for orbits that either cross, or are located closer
than the 1:2 resonance, whilst not benefiting from the pro-
tection of another mean motion resonance). In this light, we
consider it likely that the 23 day period estimate for planet c,
and any period solution longer than that, is almost certainly
a feasible, stable solution - it places that planet beyond the
location of the 1:2 mean-motion resonance, and so is stable
so long as its eccentricity is less than ∼ 0.3 (greater than this
would bring the periastron distance of planet c too close to
planet b).
The minimum possible period of 19.25 days lies interior
to the 1:2 resonance, and is close to the 3:5 resonance (period
of 19.35 days). As can be seen in the fourth row of Figure 4
in Agnew et al. (2019), this region is still likely to be stable,
so long as the orbital eccentricity for planet c is below ∼ 0.2.
5.4 Feasibility of atmospheric characterisation
Known transiting multi-planetary systems with measured
masses, around stars bright enough for atmospheric follow-
up i.e. brighter than V = 10, are exceedingly rare. The
brightness of HD 152843 (V = 8.855), combined with the
large radii of the planets, as shown in Figure 9, make them
key targets for atmospheric characterisation via transmis-
sion spectroscopy. We assess the feasibility of such an ob-
servation using the transmission spectroscopy metric (TSM;
Kempton et al. 2018), which provides the estimated SNR
of a 10 hour observation with JWST/NIRISS (Doyon et al.
2012), if a cloud-free atmosphere is assumed. Based on plan-
etary masses of 11.58 and 27.5 M⊕ (Table 3), and assuming
a mean molecular weight of 2.3, we find the TSM to be 65
and 103, for HD 152843 b and HD 152843 c, respectively.
The latter compares well with several of the targets cur-
rently included in JWST ERS and GTO programs, and is
better than the cut-off thresholds for follow-up observations,
of 96, as suggested by Kempton et al. (2018). The TSM of
103 places planet c at least amongst the top 50 % of candi-
dates suitable for atmospheric characterisation as outlined
by Kempton et al. (2018). Furthermore, as the planet mass
used to determine this value is an upper mass limit, the
TSM of planet c is likely to be significantly higher, likely
placing it amongst the top 25 % of candidates best suited
for atmospheric characterisation.
5.5 Atmospheric modelling
To assess the possibility of differentiating between differ-
ent atmospheric scenarios we generated an array of forward
models using the open source code chimera (Line et al.
2013) and compared these to synthetic observations of each
planet which were generated using PandExo (Batalha et al.
2017) for 1 transit observation using JWST NIRISS/SOSS.
A subset of these models can be seen in Figure 10. For each
planet we modelled a cloud free atmosphere with an isother-
mal temperature profile set to the derived temperature from
Table 1. For planet c we modelled the upper mass limit of
27.5 M⊕ and for planet b we considered three mass scenarios:
1) the median mass, 2) the median mass + the 3σ uncer-
tainty and 3) the median mass - the 3σ uncertainty. We
did this so that we could capture the full range of possible
transmission spectra. We then modelled the atmospheres to
have a solar C/O ratio and metalicities of 1×, 10× and 100×
solar respectively. We used the chemical grid developed by
Kreidberg et al. (2015). In Figure 10 we highlight a subset
of the models. We do not show the models for scenario 3
because the lower masses would have larger observable fea-
tures than the median and hence would be easier to observe.
For each planet we present three models: in black we show
the model for the mass and 1× solar metallicity, in pur-
ple we show the model for the mass and 100× solar and
finally in blue we show the model for the mass + 3σ and
10× metallicity. We use the mean mass and upper mass lim-
its for planets b and c, respectively. We then overplot the
predictive observations obtained from JWST NIRISS/SOSS
generated using the 1× solar median mass models. The left
panel, corresponding to planet b, shows that while with a
single transit it is possible to detect the atmosphere, there
remains a degeneracy between the metallicity and the mass
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 13
of the planet. Future RV follow-up observations will enable
us to break this degeneracy. The right panel, corresponding
to planet c, shows that the simulated data have extremely
small error bars, due to the bright star and long transit dura-
tion. These small error bars allow us to break the degeneracy
between planetary mass and atmospheric metallicity. These
simulations emphasise how promising these targets are for
follow-up measurements and atmospheric characterisation.
6 SUMMARY AND CONCLUSIONS
We present the discovery of a multi-planet system
(HD 152843, TIC 349488688, TOI 2319) with a Neptune and
a sub-Saturn sized planet, observed in Sector 25 of the nom-
inal TESS mission. The TESS light curve yields two tran-
sit events for the inner planet (Pb = 11.6264+0.0022−0.0025 days)
and a single transit event for the outer planet (Pc = 19.26-
35 days). All three transit events were identified by volun-
teers taking part in the Planet Hunters TESS citizen science
project (Eisner et al. 2020a), and the events vetted for in-
strumental and astrophysical false positives using the latte
vetting suite (Eisner et al. 2020b). Furthermore, we statisti-
cally validated both planets using the open source software
VESPA (Morton 2012, 2015b; Morton et al. 2016) by taking
into consideration the decrease in false positive probability
given the multiplicity of system (Lissauer et al. 2011, 2012,
2014).
Additionally, we obtained ground-based spectroscopic
follow-up observations with HARPS-N and EXPRES in or-
der to both constrain the orbit and planet parameters as
well as to refine the stellar properties. Joint modelling of
the light curve and RVs allowed us to constrain the mass
of the inner planet to Mb = 11.56+6.58−6.14 M⊕(2 − σ detection)
and obtain an upper mass limit for the outer planet of Mc <
27.5 M⊕. Furthermore, we constrained the orbit of the outer,
singly-transiting planet, to be between 19.26 and 35, with
the truncated model posteriors slightly favouring a period
of around 23 days. This suggests the possibility of a 2:1 res-
onance with the innermost planet.
Following this, we discuss the implications of a reso-
nance between the two planets in terms of the TTVs and
show that a 2:1 resonance would result in TTVs with an
amplitude between 5 and 40 minutes. We also show that
the planets are suitable targets for measuring the spin-
orbit alignment of the system via the RM effect, with ex-
pected amplitudes of 3.71+0.89−0.74 m s
−1 and 9.56+2.65−2.7 m s
−1 for
HD 152843 b and c, respectively.
We also show that the properties of HD 152843 c, which
likely has an extended H/He atmosphere, combined with
the brightness of the host star make it a promising tar-
gets for atmospheric characterisation. We use the TSM
(Kempton et al. 2018) to show that with a 10 hour observa-
tion with JWST/NIRISS we would obtain a SNR of 103. As
an upper mass limit was used in this calculation, the value
is likely to be significantly higher, making it a prime target
for future atmospheric characterisation.
Finally, we generate forward models of different atmo-
spheric compositions and compare these to synthetic obser-
vations for each planet in order to differentiate between dif-
ferent atmospheric scenarios. With this we show that with a
single JWST NIRISS/SOSS we would be able to detect the
atmospheres of these planets. Furthermore, the brightness of
the star combined with the transit duration of planet c re-
sults in small uncertainties in the simulated spectra, which
allow us to break the degeneracy between planetary mass
and atmospheric metallicity for the outer planet. Future RV
follow-up observations will allow us to also break this degen-
eracy for planet b.
Overall we show that this is a very promising target for
future ground and space-based follow-up observations. Con-
tinued future efforts with HARPS-N and EXPRES will be
able to conclusively determine the masses of both planets
and the orbital period of planet c, as well as search for the
RM effect. Additionally, ground-based photometers, such as
LCO/Sinistro (Brown et al. 2013), will allow us to observe
future transit events and constrain possible TTVs, as will
the space based missions such as CHEOPS (Broeg et al.
2013), or the upcoming PLATO mission (Rauer et al. 2014).
HD 152843 is also scheduled to be re-observed by the TESS
mission during Sector 52 (May-June 2022). Finally, observa-
tions with JWST or ARIEL (Tinetti et al. 2016) will help to
characterise the atmospheres of these scientifically valuable
planets.
DATA AVAILABILITY
The TESS data used within this article are hosted and made
publicly available by the Mikulski Archive for Space Tele-
scopes (MAST, http://archive.stsci.edu/tess/). Simi-
larly, the Planet Hunters TESS classifications made by the
citizen scientists can be found on the Planet Hunters Anal-
ysis Database (PHAD, https://mast.stsci.edu/phad/),
which is also hosted by MAST. The two planet can-
didates and their properties have been uploaded to
the Exoplanet Follow-up Observing Program for TESS
(ExoFOP-TESS) website as community TOIs (cTOIs;
https://exofop.ipac.caltech.edu/tess/target.php?id=349488688).
The models of the transit events and the data validation
report used for the vetting of the target were both generated
using publicly available open software codes, pyaneti and
latte.
ACKNOWLEDGEMENTS
We thank all of the volunteers who participated in the Planet
Hunters TESS project, as without them this work would not
have been possible. We also thank the editor and the ref-
eree for their comments, which improved and clarified the
manuscript. Furthermore, we are very grateful to the Di-
rector of the TNG for allocating time for the HARPS-N
observations from the directors discretionary time through
the program ID A41DDT4.
NE also thanks the LSSTC Data Science Fellowship
Program, which is funded by LSSTC, NSF Cybertraining
Grant number 1829740, the Brinson Foundation, and the
Moore Foundation; her participation in the program has
benefited this work. AM acknowledges support from the
senior Kavli Institute Fellowships. JT is a Penrose Grad-
uate Scholar and would like to thank the Oxford Physics
Endowment for Graduates (OXPEG) for funding this re-
search. Furthermore, NE, NZ, BN and SA acknowledge sup-
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
14 Eisner et al.
0 .5 0 .8 1 .0 2 .0 3 .0 5 .0
( m )
0 .0 5 0
0 .0 5 5
0 .0 6 0
0 .0 6 5
0 .0 7 0
0 .0 7 5
(R
p
/R
*)
2
[%
]
M b =  M e d ia n , [M /H ] =  1 x S ola r
M b =  M e d ia n + 3 , [M /H ] =  1 0 x S ola r
M b =  M e d ia n , [M /H ] =  1 0 0 x S ola r
N IRIS S  1  Tr a n s it
planet b
0 .5 0 .8 1 .0 2 .0 3 .0 5 .0
( m )
0 .1 5
0 .1 6
0 .1 7
0 .1 8
0 .1 9
(R
p
/R
*)
2
[%
]
M c =  Up p e r  M a ss  Lim it , [M /H ] =  1 x S ola r
M c =  Up p e r  M a ss  Lim it , [M /H ] =  1 0 x S ola r
M c =  Up p e r  M a ss  Lim it , [M /H ] =  1 0 0 x S ola r
N IRIS S  1  Tr a n s it
planet c
Figure 10. Models generated for planets b and c in the left and right panels, respectively. Each panel shows three models describing
plausible atmospheric scenarios. In black we present an atmospheric model which has a metallicity of 1× solar, considering the RV
extracted median mass and upper mass limit for planets b and c, respectively. In purple we present an atmospheric model which has
a metallicity of 100× solar considering the RV extracted median mass and upper mass limit for planets b and c, respectively. In blue
we present an atmospheric model which has a metallicity of 10× solar, however we consider the RV extracted median mass plus the 3σ
upper uncertainty for planet b and the upper mass limit for planet c. We overplot the simulated JWST NIRISS/SOSS observations for
the 1× solar case to emphasise the precision we would obtain from a single transit observation.
port from the UK Science and Technology Facilities Coun-
cil (STFC)under grant codes ST/R505006/1, ST/S505638/1
and consolidated grant no. ST/S000488. This work also re-
ceived funding from the European Research Council (ERC)
under the European UnionaˆA˘Z´s Horizon 2020 research and
innovation program (Grant agreement No. 865624).
This paper includes data collected by the TESS space-
craft,and we are grateful to the entire TESS team in ob-
taining and providing the lightcurves used in this analysis.
Funding for the TESS mission is provided by the NASA
Science Mission directorate. We obtained the publicly re-
leased TESS data from the Mikulski Archive for Space Tele-
scopes (MAST). Resources supporting this work were also
provided by the NASA High-End Computing (HEC) Pro-
gram through the NASA Advanced Supercomputing (NAS)
Division at Ames Research Center for the production of
the SPOC data products. Furthermore, these results also
made use of the Lowell Discovery Telescope at Lowell Ob-
servatory. Lowell is a private, non-profit institution dedi-
cated to astrophysical research and public appreciation of
astronomy and operates the LDT in partnership with Boston
University, the University of Maryland, the University of
Toledo, Northern Arizona University and Yale University.
This work used the EXtreme PREcision Spectrograph (EX-
PRES) that was designed and commissioned at Yale with
financial support by the U.S. National Science Foundation
under MRI-1429365 and ATI-1509436 (PI D. Fischer). Fi-
nallt, the research leading to these results has partially re-
ceived funding from the KU Leuven Research Council (grant
C16/18/005: PARADISE), from the Research Foundation
Flanders (FWO) under grant agreement G0H5416N (ERC
Runner Up Project), as well as from the BELgian federal
Science Policy Office (BELSPO) through PRODEX grant
PLATO.
Finally, NE and OB wish to thank the Asterix comics,
which provided the inspiration for our in-house nickname for
this planet system of Idefix.
This research made use of Astropy, a community-
developed core Python package for Astronomy
(Astropy Collaboration et al. 2013), matplotlib (Hunter
2007), pandas (McKinney et al. 2010), NumPy (Walt et al.
2011), astroquery (Ginsburg et al. 2019) and sklearn
(Pedregosa et al. 2011).
REFERENCES
Agnew M. T., Maddison S. T., Horner J., Kane S. R., 2019,
MNRAS, 485, 4703
Agol E., Steffen J., Sari R., Clarkson W., 2005, MNRAS, 359, 567
Aigrain S., Irwin M., 2004, MNRAS, 350, 331
Akeson R. L., et al., 2013, PASP, 125, 989
Anglada-Escude´ G., Butler R. P., 2012, ApJS, 200, 15
Astropy Collaboration et al., 2013, A&A, 558, A33
Bailer-Jones C. A. L., Rybizki J., Fouesneau M., Mantelet G.,
Andrae R., 2018, AJ, 156, 58
Baranne A., et al., 1996, A&AS, 119, 373
Barraga´n O., Gandolfi D., Antoniciello G., 2019, MNRAS,
482, 1017
Barros S. C. C., et al., 2017, A&A, 608, A25
Batalha N. E., et al., 2017, PASP, 129, 064501
Batalha N. E., Lewis T., Fortney J. J., Batalha N. M., Kempton
E., Lewis N. K., Line M. R., 2019, ApJ, 885, L25
Blackman R. T., et al., 2020, AJ, 159, 238
Borucki W. J., Koch D., Kepler Science Team 2010, in
AAS/Division for Planetary Sciences Meeting Abstracts #42.
p. 47.03
Brakensiek J., Ragozzine D., 2016, The Astrophysical Journal,
821, 47
Broeg C., et al., 2013, in EPJ Web of Conferences. p. 03005
Brown T. M., et al., 2013, PASP, 125, 1031
Bruntt H., et al., 2010, MNRAS, 405, 1907
Buchhave L. A., et al., 2012, Nature, 486, 375
Buchhave L. A., et al., 2014, Nature, 509, 593
Carleo I., et al., 2020, AJ, 160, 114
Chatterjee S., Ford E. B., Matsumura S., Rasio F. A., 2008, ApJ,
686, 580
Chen H., Rogers L. A., 2016, ApJ, 831, 180
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 15
Choi J., Dotter A., Conroy C., Cantiello M., Paxton B., Johnson
B. D., 2016, ApJ, 823, 102
Christiansen J. L., et al., 2017, AJ, 154, 122
Claret A., 2016, VizieR Online Data Catalog,
pp J/A+A/600/A30
Cosentino R., et al., 2012, in McLean I. S., Ramsay S. K.,
Takami H., eds, Society of Photo-Optical Instrumentation
Engineers (SPIE) Conference Series Vol. 8446, Ground-based
and Airborne Instrumentation for Astronomy IV. p. 84461V,
doi:10.1117/12.925738
Cosentino R., et al., 2014, in Ramsay S. K., McLean I. S.,
Takami H., eds, Society of Photo-Optical Instrumentation
Engineers (SPIE) Conference Series Vol. 9147, Ground-based
and Airborne Instrumentation for Astronomy V. p. 91478C,
doi:10.1117/12.2055813
Crossfield I. J. M., et al., 2016, ApJS, 226, 7
Cutri R. M., et al. 2013, VizieR Online Data Catalog, p. II/328
Cutri R. M., et al., 2003, VizieR Online Data Catalog, p. II/246
Dalal S., Hebrard G., Lecavelier Des Etangs A., Petit A., Bour-
rier V., Laskar J., Konig P.-C., Correia A. C. M., 2019, in
AAS/Division for Extreme Solar Systems Abstracts. p. 202.04
Deck K. M., Agol E., Holman M. J., Nesvorny´ D., 2014, ApJ,
787, 132
Deeg H. J., et al., 2009, Astronomy and Astrophysics, 506, 343
Dietrich J., Apai D., 2020, AJ, 160, 107
Doyon R., et al., 2012, in Space Telescopes and Instrumentation
2012: Optical, Infrared, and Millimeter Wave. p. 84422R
Dragomir D., et al., 2019a, The Astrophysical Journal Letters,
875, L7
Dragomir D., et al., 2019b, ApJ, 875, L7
Dressing C. D., et al., 2015, ApJ, 800, 135
Eisner N. L., et al., 2020a, arXiv e-prints, p. arXiv:2011.13944
Eisner N., Lintott C., Aigrain S., 2020b,
Journal of Open Source Software, 5, 2101
Eisner N. L., et al., 2020c, MNRAS, 494, 750
Espinoza N., et al., 2020, MNRAS, 491, 2982
Feroz F., Hobson M. P., Cameron E., Pettitt A. N., 2019,
The Open Journal of Astrophysics, 2, 10
Findeisen K., Hillenbrand L., Soderblom D., 2011, AJ, 142, 23
Fischer D. A., et al., 2012, MNRAS, 419, 2900
Fortenbach C. D., Dressing C. D., 2020, PASP, 132, 054501
Fridlund M., et al., 2020, MNRAS, 498, 4503
Gaia Collaboration 2018, VizieR Online Data Catalog, p. I/345
Gaia Collaboration Brown A. G. A., Vallenari A., Prusti T., de
Bruijne J. H. J., Babusiaux C., Biermann M., 2020, arXiv
e-prints, p. arXiv:2012.01533
Gaia C., et al., 2018, Astronomy & Astrophysics, 616
Gandolfi D., et al., 2019b, ApJ, 876, L24
Gandolfi D., et al., 2019a, The Astrophysical Journal Letters, 876,
L24
Gao P., Zhang X., 2020, ApJ, 890, 93
Gilbert E. A., et al., 2020, AJ, 160, 116
Ginsburg A., et al., 2019, The Astronomical Journal, 157, 98
Girardi L., Groenewegen M. A. T., Hatziminaoglou E., da Costa
L., 2005, A&A, 436, 895
Hjorth M., Albrecht S., Hirano T., Winn J. N., Daw-
son R. I., Zanazzi J. J., Knudstrup E., Sato B.,
2021, Proceedings of the National Academy of Science,
118, 2017418118
Høg E., et al., 2000, Astronomy and Astrophysics, 355, L27
Holman M. J., Murray N. W., 2005, Science, 307, 1288
Horner J., et al., 2019, AJ, 158, 100
Huang C. X., et al., 2018, ApJ, 868, L39
Hunter J. D., 2007, Computing in Science Engineering, 9, 90
Jenkins J. M., et al., 2016, in Software and Cyberinfrastructure
for Astronomy IV. p. 99133E, doi:10.1117/12.2233418
Jurgenson C., Fischer D., McCracken T., Sawyer D., Szymkowiak
A., Davis A., Muller G., Santoro F., 2016, in Evans C. J.,
Simard L., Takami H., eds, Society of Photo-Optical In-
strumentation Engineers (SPIE) Conference Series Vol. 9908,
Ground-based and Airborne Instrumentation for Astronomy
VI. p. 99086T (arXiv:1606.04413), doi:10.1117/12.2233002
Kempton E. M. R., et al., 2018, PASP, 130, 114401
Kipping D. M., 2013, MNRAS, 435, 2152
Kova´cs G., Zucker S., Mazeh T., 2002, A&A, 391, 369
Kreidberg L., 2015, PASP, 127, 1161
Kreidberg L., et al., 2015, ApJ, 814, 66
Kurucz R.-L., 1993, Kurucz CD-Rom, 13
Latham D. W., et al., 2011, ApJ, 732, L24
Lee E. J., Chiang E., 2016, ApJ, 817, 90
Leleu A., et al., 2021, arXiv e-prints, p. arXiv:2101.09260
Levine S. E., et al., 2012, in Ground-based and Airborne Tele-
scopes IV. p. 844419
Lienhard F., et al., 2020, MNRAS, 497, 3790
Line M. R., et al., 2013, ApJ, 775, 137
Lintott C. J., et al., 2008, MNRAS, 389, 1179
Lintott C., et al., 2011, VizieR Online Data Catalog,
p. J/MNRAS/410/166
Lissauer J. J., et al., 2011, ApJS, 197, 8
Lissauer J. J., et al., 2012, ApJ, 750, 112
Lissauer J. J., et al., 2014, ApJ, 784, 44
Mamajek E. E., Hillenbrand L. A., 2008, ApJ, 687, 1264
Mandel K., Agol E., 2002, Astrophysical Journal, 580, L171
Mann A. W., et al., 2020, AJ, 160, 179
McKinney W., et al., 2010, in Proceedings of the 9th Python in
Science Conference. pp 51–56
McLaughlin D. B., 1924, ApJ, 60, 22
Millholland S., 2019, ApJ, 886, 72
Mortier A., Sousa S. G., Adibekyan V. Z., Branda˜o I. M., Santos
N. C., 2014, A&A, 572, A95
Mortier A., et al., 2020, MNRAS, 499, 5004
Morton T. D., 2012, ApJ, 761, 6
Morton T. D., 2015a, VESPA: False positive probabilities calcu-
lator, Astrophysics Source Code Library (ascl:1503.011)
Morton T. D., 2015b, VESPA: False positive probabilities calcu-
lator, Astrophysics Source Code Library (ascl:1503.011)
Morton T. D., Bryson S. T., Coughlin J. L., Rowe J. F., Ravichan-
dran G., Petigura E. A., Haas M. R., Batalha N. M., 2016,
ApJ, 822, 86
Noyes R. W., Hartmann L. W., Baliunas S. L., Duncan D. K.,
Vaughan A. H., 1984, ApJ, 279, 763
Osborn H. P., et al., 2016, MNRAS, 457, 2273
Owen J. E., Wu Y., 2016, ApJ, 817, 107
Pearson K. A., Palafox L., Griffith C. A., 2018, MNRAS, 474, 478
Pedregosa F., et al., 2011, the Journal of machine Learning re-
search, 12, 2825
Petersburg R. R., et al., 2020, AJ, 159, 187
Queloz D., et al., 2010, A&A, 517, L1
Quinn S. N., et al., 2019, The Astronomical Journal, 158, 177
Rauer H., et al., 2014, Experimental Astronomy, 38, 249
Ricker G. R., et al., 2015, Journal of Astronomical Telescopes, Instruments, and Systems,
1, 014003
Rossiter R. A., 1924, ApJ, 60, 15
Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525
Schneider J., 2000, in From Giant Planets to Cool Stars. p. 284
Shulyak D., Tsymbal V., Ryabchikova T., Stu¨tz C., Weiss W. W.,
2004, A&A, 428, 993
Sneden C. A., 1973, PhD thesis, THE UNIVERSITY OF TEXAS
AT AUSTIN.
Sousa S. G., 2014, ARES + MOOG: A Practical Overview of an
Equivalent Width (EW) Method to Derive Stellar Parame-
ters. pp 297–310, doi:10.1007/978-3-319-06956-2 26
Sousa S. G., Santos N. C., Adibekyan V., Delgado-Mena E., Is-
raelian G., 2015, A&A, 577, A67
Stassun K. G., Torres G., 2016, AJ, 152, 180
Stassun K. G., Torres G., 2018, ApJ, 862, 61
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
16 Eisner et al.
Stassun K. G., Collins K. A., Gaudi B. S., 2017, AJ, 153, 136
Stassun K. G., Corsaro E., Pepper J. A., Gaudi B. S., 2018, AJ,
155, 22
Stassun K. G., et al., 2019, arXiv e-prints, p. arXiv:1905.10694
Storch N. I., Lai D., Anderson K. R., 2017,
Monthly Notices of the Royal Astronomical Society,
465, 3927
Teske J., et al., 2020, AJ, 160, 96
Tinetti G., et al., 2016, in MacEwen H. A., Fazio G. G.,
Lystrup M., Batalha N., Siegler N., Tong E. C., eds, Soci-
ety of Photo-Optical Instrumentation Engineers (SPIE) Con-
ference Series Vol. 9904, Space Telescopes and Instrumenta-
tion 2016: Optical, Infrared, and Millimeter Wave. p. 99041X,
doi:10.1117/12.2232370
Tkachenko A., 2015, A&A, 581, A129
Torres G., Andersen J., Gime´nez A., 2010, The Astronomy and
Astrophysics Review, 18, 67
Tremaine S., Dong S., 2012, AJ, 143, 94
Trifonov T., Rybizki J., Ku¨rster M., 2019, A&A, 622, L7
Vanderburg A., et al., 2019, ApJ, 881, L19
Walt S. v. d., Colbert S. C., Varoquaux G., 2011, Computing in
science & engineering, 13, 22
Wang L., Dai F., 2019, ApJ, 873, L1
Winn J. N., 2010, Exoplanet Transits and Occultations. Univer-
sity of Arizona Press, pp 55–77
Yee S. W., et al., 2018, AJ, 155, 255
Zucker S., Giryes R., 2018, AJ, 155, 147
APPENDIX A: CONFIRMED MULTI-PLANET
SYSTEMS
This paper has been typeset from a TEX/LATEX file prepared by
the author.
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021
O
R
IG
IN
A
L
 U
N
E
D
IT
E
D
 M
A
N
U
S
C
R
IP
T
HD152843 17
Table A1. Bright multi-planet system.
Host name Planet letter Rpl (R⊕) Mpl (M⊕) Ppl (days) Vmag No. confirmed planets Reference
GJ 143 b 2.610.17−0.16 22.7
2.2
−1.9 35.61253
0.0006
−0.00062 8.08 2 Dragomir et al. (2019b)
HAT-P-11 b 4.360.06−0.06 26.698
2.22
−2.22 4.8878 9.46 2 Yee et al. (2018)
HD 106315 b 2.440.17−0.17 12.6
3.2
−3.2 9.55237
0.00089
−0.00089 8.951 2 Barros et al. (2017)
c 4.350.23−0.23 15.2
3.7
−3.7 21.05704
0.00046
−0.00046
HD 15337 b 1.640.06−0.06 7.51
1.09
−1.01 4.75615
0.00017
−0.00017 9.1 2 Gandolfi et al. (2019b)
c 2.390.12−0.12 8.11
1.82
−1.69 17.1784
0.0016
−0.0016
HD 213885 b 1.7450.05−0.05 8.83
0.66
−0.65 1.00804
0.00002
−0.00002 7.95 2 Espinoza et al. (2020)
HD 23472 b 1.8721.32−1.32 17.92
1.41
−14.0 17.667
0.142
−0.095 9.73 2 Trifonov et al. (2019)
c 2.1490.34−0.34 17.18
1.07
−13.77 29.625
0.224
−0.171
HD 3167 b 1.70.08−0.08 5.02
0.38
−0.38 0.95962
0.00003
−0.00003 8.97 3 Christiansen et al. (2017)
c 2.860.22−0.22 9.8
1.3
−1.24 29.83832
0.00291
−0.0032
HD 39091 c 2.0420.05−0.05 4.82
0.84
−0.86 6.2679
0.00046
−0.00046 5.65 2 Huang et al. (2018)
HD 86226 c 2.160.08−0.08 7.25
1.19
−1.12 3.98442
0.00018
−0.00018 7.93 2 Teske et al. (2020)
Kepler-93 b 1.5690.11−0.11 4.544
0.85
−0.85 4.72674
0.000001
−0.000001 9.996 2 Dressing et al. (2015)
TOI-421 b 2.680.19−0.18 7.17
0.66
−0.66 5.19672
0.00049
−0.00049 9.931 2 Carleo et al. (2020)
c 5.090.16−0.15 16.42
1.06
−1.04 16.06819
0.00035
−0.00035
WASP-8 b 12.6660.56−0.56 807.288
104.88
−104.88 8.15872
0.00001
−0.00001 9.789 2 Queloz et al. (2010)
Note – Confirmed exoplanets from the NASA Exoplanet Archive that are members of systems with more than one planet, with mass
measurements better than 50 % and around stars brighter than V = 10 (Akeson et al. 2013). All parameters are as listed in the
NASA Exoplanet Archive as of April 2021.
D
ow
nloaded from
 https://academ
ic.oup.com
/m
nras/advance-article/doi/10.1093/m
nras/stab1253/6274703 by U
niversity of C
am
bridge user on 01 June 2021