mdevora commited on
Commit
9c5a860
·
1 Parent(s): d47f496

Upload 19 files

Browse files
.gitattributes CHANGED
@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ documents/txt/armstrong2020.pdf filter=lfs diff=lfs merge=lfs -text
37
+ documents/txt/Giacalone_2021_AJ_161_24.pdf filter=lfs diff=lfs merge=lfs -text
38
+ documents/txt/Valizadegan_2022_ApJ_926_120.pdf filter=lfs diff=lfs merge=lfs -text
documents/txt/Giacalone_2021_AJ_161_24.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ced790bd36c3c285049b8ba2f67e5e5647d9c75f41895f9c1f68c00a688dbba2
3
+ size 3062765
documents/txt/Valizadegan_2022_ApJ_926_120.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4658a3b3579ec40cda503a3a6be40d497e38c711fe9d1900dcf931871abcb85f
3
+ size 6608266
documents/txt/antoine.txt ADDED
@@ -0,0 +1,1678 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ A&A 664, A113 (2022)
2
+ https://doi.org/10.1051/0004-6361/202243554
3
+ © A. Thuillier et al. 2022
4
+
5
+ A search for transiting planets around hot subdwarfs
6
+
7
+ II. Supplementary methods and results from TESS Cycle 1
8
+
9
+ A. Thuillier1,2
10
+
11
+ , V. Van Grootel1, M. Dévora-Pajares3, F. J. Pozuelos1,4, S. Charpinet5, and L. Siess2
12
+
13
+ 1 Space sciences, Technologies and Astrophysics Research (STAR) Institute, Université de Liège, 19C Allée du 6 Août, 4000 Liège,
14
+
15
+ Belgium
16
+ e-mail: antoine.thuillier@uliege.be
17
+
18
+ 2 Institut d’Astronomie et d’Astrophysique, Université Libre de Bruxelles (ULB), CP 226 1050 Bruxelles, Belgium
19
+
20
+ e-mail: antoine.thuillier@ulb.be
21
+
22
+ 3 Dpto. Física Teórica y del Cosmos, Universidad de Granada, 18071, Granada, Spain
23
+ 4 Astrobiology Research Unit, Université de Liège, Allée du 6 Août 19C, 4000 Liège, Belgium
24
+ 5 Institut de Recherche en Astrophysique et Planétologie, CNRS, Université de Toulouse, CNES, 14 avenue Edouard Belin, 31400
25
+
26
+ Toulouse, France
27
+
28
+ Received 15 March 2022 / Accepted 25 May 2022
29
+
30
+ ABSTRACT
31
+
32
+ Context. Hot subdwarfs, which are hot and small He-burning objects, are ideal targets for exploring the evolution of planetary systems
33
+ after the red giant branch (RGB). Thus far, no planets have been confirmed around them, and no systematic survey to find planets has
34
+ been carried out.
35
+ Aims. In this project, we aim to perform a systematic transit survey in all light curves of hot subdwarfs from space-based telescopes
36
+ (Kepler, K2, TESS, and CHEOPS). The goal is to compute meaningful statistics on two points: firstly, the occurrence rates of planets
37
+ around hot subdwarfs, and secondly, the probability of survival for close-in planets engulfed during the RGB phase of their host. This
38
+ paper focuses on the analysis of the observations carried out during cycle 1 of the TESS mission.
39
+ Methods. We used our specifically designed pipeline SHERLOCK to search for transits in the available light curves. When a signal is
40
+ detected, it is processed in the next evaluating stages before an object is qualified for follow-up observations and in-depth analysis to
41
+ determine the nature of the transiting body.
42
+ Results. We applied our method to the 792 hot subdwarfs observed during cycle 1 of TESS. While 378 interesting signals were
43
+ detected in the light curves, only 26 stars were assigned for follow-up observations. We have identified a series of eclipsing binaries,
44
+ transiting white dwarfs, and other types of false positives, but no planet has been confirmed thus far. A first computation of the upper
45
+ limit for occurrence rates was made with the 549 targets displaying no signal.
46
+ Conclusions. The tools and method we developed proved their efficiency in analysing the available light curves from space missions,
47
+ from detecting an interesting signal to identifying a transiting planet. This will allow us to fulfil the two main goals of this project.
48
+
49
+ Key words. planet-star interactions – planetary systems – stars: horizontal-branch – subdwarfs – techniques: photometric
50
+
51
+ 1. Introduction
52
+
53
+ Hot subdwarfs are evolved and compact stars coming in two
54
+ main flavours, subdwarfs of type B (sdB) and subdwarfs of type
55
+ O (sdO). The sdB stars have Teff = 20 000–40 000 K and log g =
56
+ 5.2–6.2 (Saffer et al. 1994; Green et al. 2008), while sdO stars are
57
+ hotter (Teff = 40 000–80 000 K), and have a wide range of sur-
58
+ face gravities (log g = 4.0–6.5; Oreiro Rey et al. 2004; Johnson
59
+ et al. 2014). The sdB stars lie on the blue tail of the horizontal
60
+ branch, the so-called extreme horizontal branch (EHB), which
61
+ identifies them as core-He burning stars (Heber 1986). They have
62
+ lost most of their envelope during the ascension of the first red
63
+ giant branch (RGB), and they now have extremely thin resid-
64
+ ual H envelopes (Menv < 0.01 M⊙, Heber 1986). This extremely
65
+ thin envelope explains the atmospheric parameters of sdB stars
66
+ and their inability to sustain H-shell burning, and prevents them
67
+ from ascending the asymptotic giant branch (AGB) after core-He
68
+ exhaustion (Dorman et al. 1993). They rather directly evolve to
69
+ the white dwarf stage through the sdO type during the immediate
70
+ post-EHB phase.
71
+
72
+ In addition to the direct progenies of sdB stars, the com-
73
+ pact sdO stars (log g = 5.2–6.5) could also be direct post-RGB
74
+ objects through a so-called late hot He-flash (Miller Bertolami
75
+ et al. 2008), or they might be end products of merger events
76
+ (Webbink 1984; Iben & Tutukov 1984; Iben 1990; Saio & Jeffery
77
+ 2000, 2002). Accordingly, these stars do not descend from the
78
+ sdB stars. sdO stars with log g < 5.2 exist as well, which are
79
+ post-AGB stars, that is, stars that have ascended the giant branch
80
+ a second time after core-He burning exhaustion (Reindl et al.
81
+ 2016). We are not interested in these in this study because we are
82
+ focusing on post-RGB hot subdwarfs. An in-depth review of hot
83
+ subdwarfs can be found in Heber (2016).
84
+
85
+ There are several planet candidates around hot subdwarfs.
86
+ They were identified through various methods such as reflec-
87
+ tion signals in Kepler light curves (Charpinet et al. 2011; Bear &
88
+ Soker 2014) or stellar pulsation timing variations (Silvotti et al.
89
+ 2007). A review of the search for planets around hot subdwarfs
90
+ can be found in Van Grootel et al. (2021). None of these can-
91
+ didates have been confirmed to date; some are heavily debated
92
+ (Krzesinski 2015; Blokesz et al. 2019), and others were discarded
93
+
94
+ Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0),
95
+ which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
96
+ This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication.
97
+
98
+ A113, page 1 of 12
99
+
100
+ Astronomy&Astrophysics A&A 664, A113 (2022)
101
+
102
+ later (Silvotti et al. 2018; Krzesinski et al. 2020). A mini radial
103
+ velocity (RV) survey carried out with the HARPS-N spectro-
104
+ graph of eight apparently single hot subdwarfs gave null results
105
+ down to a few Jupiter masses (Silvotti et al. 2020). More gen-
106
+ erally, in regard to horizontal branch (HB) stars, a single planet
107
+ around an F2-type HB star was announced by Setiawan et al.
108
+ (2010), but was later discarded by Jones & Jenkins (2014). The
109
+ transit method has never been used at a large scale to search for
110
+ planets transiting hot subdwarfs.
111
+
112
+ Due to the loss of most of their envelope during the RGB,
113
+ hot subdwarfs are small stars (0.1–0.3 R⊙; Heber 2016). This
114
+ property makes them ideal targets for using the transit method
115
+ to address the question of the evolution of exoplanetary systems
116
+ directly after the RGB phase of evolution. We indeed aim in this
117
+ project to perform a transit survey in all available light curves
118
+ of hot subdwarfs from space-based telescopes. The main objec-
119
+ tive is to determine the occurrence rates of planets around these
120
+ stars as a function of orbital period and planetary radius. We
121
+ also aim to place strong observational constraints on the sur-
122
+ vival of close-in planets that were engulfed during the RGB
123
+ phase of their host star, which are currently completely missing.
124
+ “Close-in planets here mean those with current orbital periods
125
+ up to ∼ 50 days (orbital radii up to 0.20 AU), which have the
126
+ highest transit probability (which is about 0.35% at 50 days
127
+ orbital period). Considering main-sequence masses between 1
128
+ and ∼2.5 M⊙ that are the main hot subdwarf progenitors, all these
129
+ planets could have been engulfed during the RGB phase of the
130
+ host star (Villaver & Livio 2009).
131
+
132
+ This paper is the second of the series started by Van Grootel
133
+ et al. (2021), where the context of the project was introduced,
134
+ together with the results from injection-and-recovery tests on
135
+ actual light curves from the Kepler (Borucki et al. 2010), K2
136
+ (Howell et al. 2014), and TESS (Ricker et al. 2014) space
137
+ missions. From these performance tests, we determined which
138
+ transiting bodies in terms of object radius and orbital period
139
+ we are able to detect with our tools. In TESS data, we are able
140
+ to detect transiting bodies of ≲2 R⊕ with orbital periods shorter
141
+ than 15 days for a magnitude (G mag) between 13 and 14 and a
142
+ single sector (27 days) of observations (Van Grootel et al. 2021).
143
+ For brighter stars and/or stars that were observed in more sectors
144
+ (up to 13 for the primary mission, which represents one year of
145
+ continuous observation), we are able to detect bodies with radii
146
+ smaller than 1 R⊕, as well as bodies with radii of a few R⊕ for
147
+ orbital periods longer than 35 days (see details in Table 3 of Van
148
+ Grootel et al. 2021).
149
+
150
+ This second paper focuses on the full explanation of the
151
+ light-curve analysis process and on the first results obtained for
152
+ the hot subdwarfs observed by TESS during its cycle 1. Section 2
153
+ presents the data we used and the sample selection process, fol-
154
+ lowed by a description of the method we used for the transit
155
+ survey along with a brief description of our main tools. We also
156
+ detail the different steps of the analysis and how we ranked the
157
+ detected signals. In Sect. 3, we present the results and detail the
158
+ number of signals in each stage of the analysis. We also report
159
+ some examples of false positives that we encountered during our
160
+ survey. We conclude in Sect. 4 with a discussion and describe
161
+ the next steps of the project.
162
+
163
+ 2. Data and method
164
+
165
+ 2.1. Data
166
+
167
+ TESS observed 1302 sdB and sdO stars in 2-min short-cadence
168
+ (SC) mode during its primary mission (Van Grootel et al. 2021).
169
+
170
+ A113, page 2 of 12
171
+
172
+ Table 1. Detailed statistics for hot subdwarfs observed in SC mode
173
+ during the primary mission of TESS (July 2018–July 2020).
174
+
175
+ Number of
176
+ sectors
177
+
178
+ Primary Cycle Cycle
179
+ 1
180
+ mission
181
+
182
+ 2
183
+
184
+ 1
185
+ 2
186
+ 3
187
+ 4
188
+ 5
189
+ 6
190
+ 7
191
+ 8
192
+ 9
193
+ 10
194
+ 11
195
+ 12
196
+ 13
197
+
198
+ 877
199
+ 205
200
+ 72
201
+ 23
202
+ 21
203
+ 24
204
+ 7
205
+ 10
206
+ 6
207
+ 6
208
+ 13
209
+ 23
210
+ 15
211
+
212
+ Total
213
+
214
+ Mean sect./star
215
+
216
+ 1302
217
+
218
+ 2.1
219
+
220
+ 627
221
+ 95
222
+ 25
223
+ 7
224
+ 3
225
+ 10
226
+ 2
227
+ 5
228
+ 3
229
+ 1
230
+ 3
231
+ 7
232
+ 4
233
+
234
+ 792
235
+
236
+ 1.6
237
+
238
+ 250
239
+ 110
240
+ 47
241
+ 16
242
+ 18
243
+ 14
244
+ 5
245
+ 5
246
+ 3
247
+ 5
248
+ 10
249
+ 16
250
+ 11
251
+
252
+ 510
253
+
254
+ 2.8
255
+
256
+ Notes. Similar to Table 1 from Van Grootel et al. (2021), but with
257
+ details for both cycles.
258
+
259
+ Two so-called cycles compose this primary mission. Cycle 1
260
+ (July 2018–July 2019) observed the southern and cycle 2 (July
261
+ 2019–July 2020) the northern celestial hemisphere. The two
262
+ hemispheres are revisited during the extended mission (July
263
+ 2020–July 2022), with cycle 3 (southern) and cycle 4 (northern).
264
+ Except for cycle 4, which has 16 sectors, each cycle is divided
265
+ into 13 sectors of ∼27 days of observations, corresponding to
266
+ two orbits of the satellite.
267
+
268
+ There is a significant disparity in the length of the obser-
269
+ vations available from target to target. The TESS sectors overlap
270
+ near the celestial poles, which means that a few targets have been
271
+ observed almost continuously during one year, while the large
272
+ majority are visible in a single sector of only ∼27 days. Table 1
273
+ displays the detailed statistics for hot subdwarfs for cycles 1
274
+ and 2 of the primary mission. Interestingly, a difference appears
275
+ between the two cycles: the number of stars in one single sector
276
+ in cycle 1 is significantly larger than in cycle 2. This is striking
277
+ when the mean number of sectors per star is computed, which
278
+ is 1.6 in cycle 1 and 2.8 in cycle 2. This discrepancy is directly
279
+ linked to the target positions in the sky (see Fig. 1 from Van
280
+ Grootel et al. 2021), which in turn is most likely linked to our
281
+ position in the Galaxy and our height above the Galactic plane.
282
+ We focus in this paper on the hot subdwarfs observed in
283
+ TESS cycle 1, completed by cycle 3 in case of detection of an
284
+ interesting signal (see below). Cycle 1 contains sectors 1–13,
285
+ while cycle 3 contains sectors 27–39. Cycle 2 (sectors 14–26)
286
+ and cycle 4 (sectors 40–56) will be explored in the next steps
287
+ of this project. TESS provides flux measurements with differ-
288
+ ent exposure times. The work presented here mainly used the
289
+ 2-minute SC mode. In its primary mission, TESS only provided
290
+ data with cadences of 2 and 30 min, while in its extended mis-
291
+ sion, two additional observation modes are available at 20-s and
292
+ 10-min cadences. We occasionally used these other cadences
293
+ (20 s, 10 min, and 30 min) to probe interesting signals detected
294
+ in the 2-min mode.
295
+
296
+ The data used in this study are the pre-search data condition-
297
+ ing simple aperture photometry (PDC-SAP) light curves, which
298
+
299
+ A. Thuillier et al.: A transit survey to search for planets around hot subdwarfs. II.
300
+
301
+ are the sum of the flux from the pixels in a pre-defined aperture
302
+ corrected for long-term trends. They were directly downloaded
303
+ from the NASA Mikulski Archive for Space Telescopes (MAST)
304
+ database. Other data were gathered from the TESS Asteroseis-
305
+ mic Science Operation Center, for example the identification of
306
+ the pulsating hot subdwarfs.
307
+
308
+ 2.2. Light-curve analysis: Method and tools
309
+
310
+ 2.2.1. SHERLOCK pipeline
311
+
312
+ The main analysis was conducted using our pipeline Search-
313
+ ing for Hints of Exoplanets fRom Light curves Of spaCe-based
314
+ seeKers SHERLOCK1 (Pozuelos et al. 2020), which is a ver-
315
+ satile tool we recently developed to detect shallow periodic
316
+ transits in light curves from space-based observatories such as
317
+ Kepler/K2 and TESS. SHERLOCK has six different modules that
318
+ allow (1) downloading and preparing the light curves from their
319
+ online repositories, (2) searching for planetary candidates, (3)
320
+ performing a semi-automatic vetting of the interesting signals,
321
+ (4) computing a statistical validation, (5) modeling the signals
322
+ to refine their ephemerides, and (6) computing observational
323
+ windows from ground-based observatories to trigger a follow-
324
+ up campaign. In addition, to optimize the planetary search,
325
+ SHERLOCK executes an automatic process that allows for initial
326
+ corrections such as masking high-noise regions and correcting
327
+ strong variability caused by fast rotators. Then, SHERLOCK per-
328
+ forms a user-defined multi-detrend approach using the bi-weight
329
+ method (Hippke et al. 2019) by varying the window size a num-
330
+ ber of times. In our case, we used 12 different detrends. Then,
331
+ SHERLOCK searches for planetary candidates in the original
332
+ PDC-SAP flux jointly with the 12 newly detrended light curves.
333
+ This strategy allows us to search for the most appropriate detrend
334
+ maximising the signal-to-noise ratio (S/N) and signal detection
335
+ efficiency (SDE)2 of any detection. The six modules are imple-
336
+ mented so that the user only needs to execute a few lines of code,
337
+ which avoides diving into many different pipelines, codes, and
338
+ processes. Moreover, SHERLOCK has direct access to short- and
339
+ long-cadence data observed by Kepler/K2 and TESS.
340
+
341
+ A typical SHERLOCK execution starts with the automatic
342
+ download of the PDC-SAP flux light curve of the desired tar-
343
+ get from the MAST. Depending on the options that are set, it is
344
+ possible to analyse all the available data of the given target or
345
+ use only specified sectors (for TESS), quarters (Kepler), or cam-
346
+ paigns (K2). Then, after the initial corrections described above,
347
+ the planetary search starts using a modified version of the transit
348
+ least- squares (TLS) package (Hippke & Heller 2019). This mod-
349
+ ified version consists of a proper reduction of the period-range
350
+ density over which TLS searches for planets. This reduction
351
+ allows us to avoid extremely dense searches due to observa-
352
+ tions performed over sectors with large time gaps between them,
353
+ which translates into a much faster but less sensitive execu-
354
+ tion. When SHERLOCK spots a periodic signal with S/N and
355
+ SDE above the minimum thresholds defined by the user, in
356
+ our case, 6 and 8, respectively, this signal is masked, and the
357
+ search is repeated until no other signals above the thresholds
358
+ are found. This search-and-mask process is called “run”. In this
359
+ study, we set a limit of six runs to avoid wasting computational
360
+ time while maximising the chances to detect transits. After six
361
+ runs, the light curves have many masked regions coming from
362
+
363
+ 1 SHERLOCK’s code is open-source, user-friendly, and available on
364
+ GitHub: https://github.com/franpoz/SHERLOCK
365
+ 2 SDE is computed as 1−⟨S R⟩
366
+ σ(S R)
367
+ a tested period, and σ is the standard deviation.
368
+
369
+ , with SR defined as the signal residue for
370
+
371
+ Fig. 1. SHERLOCK output of the best result of the first run of TIC
372
+ 142875987 in sectors 4 (panels 1 to 3, from top to bottom) and 31 (pan-
373
+ els 4 to 6) of TESS. The secondary eclipse is clearly visible in the top
374
+ panel of both sectors (panels 1 and 4) where the detrended light curve
375
+ (grey) and position of the detected transits (red) are shown. Panels 2
376
+ and 5: phase-folded light curve over the period of the spotted transit.
377
+ Panels 3 and 6: power spectrum with the main signal and its harmonics
378
+ highlighted.
379
+
380
+ the previous findings, implying that any new detection will be
381
+ affected by these gaps. This reduces their credibility and there-
382
+ fore makes the following validation checks more challenging. All
383
+ the information regarding the applied detrending, initial mask-
384
+ ing, fast rotation correction, and planetary searches are stored
385
+ automatically in folders and log files. Then, the user needs to
386
+ inspect the results to verify the findings visually. An example of
387
+ a SHERLOCK finding is shown in Fig. 1. The figure shows two
388
+ different observations of TIC 142875987, one during cycle 1 (top
389
+ three panels), and the other from cycle 3 (bottom three panels).
390
+ Panels 1 and 4 (from top to bottom) show the light curve with
391
+ the position of the spotted transits, panels 2 and 5 display the
392
+
393
+ A113, page 3 of 12
394
+
395
+ inFig. A&A 664, A113 (2022)
396
+
397
+ phase light curve folded over the period of the detected signal,
398
+ while panels 3 and 6 show the power spectrum of the phase-
399
+ folded light curve for all the possible periods, with highlights on
400
+ the selected signal and its harmonics.
401
+
402
+ 2.2.2. FELIX
403
+
404
+ Some of the light curves display too much variability, typically
405
+ related to stellar oscillations or to the presence of a companion
406
+ star, to be directly analysed by the SHERLOCK pipeline. In these
407
+ cases, we used FELIX (Charpinet et al. 2010; Zong et al. 2016),
408
+ a tool designed to extract interactively or automatically periodic
409
+ variations in a light curve. FELIX subtracts from the light curve
410
+ each periodic variation that is spotted above a pre-defined thresh-
411
+ old (usually the threshold corresponding to a 4σ significance)
412
+ using the pre-withening technique (Deeming 1975). That is, we
413
+ identified the frequency and amplitude of the highest-amplitude
414
+ peak in the Lomb–Scargle Periodogram (LSP) of the light curve.
415
+ They were used as initial guesses in a subsequent non-linear
416
+ least-squares (NLLS) fit of a cosine wave in time domain using
417
+ the Levenberg–Marquardt algorithm. The fitted wave of the
418
+ derived frequency, amplitude, and phase was then subtracted
419
+ from the light curve. The operation was repeated as long as
420
+ there was a peak above the pre-defined threshold. The light curve
421
+ cleaned from stellar oscillations and other periodic variabilities
422
+ (e.g. linked to binarity) was then given to SHERLOCK to start
423
+ the transit search. This cleaning by FELIX improved our capa-
424
+ bility to detect small planets in light curves dominated by stellar
425
+ oscillations or other periodic variabilities. An example is given
426
+ in Fig. 2, which presents the case of the hybrid sdB pulsator
427
+ TIC 169285097. The upper panel presents the light curve (SC
428
+ cadence, sector 2) before (original) and after the pre-withening
429
+ process, that is, after the oscillations were removed. The two bot-
430
+ tom panels show the results of injection-and-recovery tests per-
431
+ formed with our Multi-phAse Transits Recovery from Injected
432
+ eXoplanets MATRIX3 tool (Dévora-Pajares & Pozuelos 2022),
433
+ on the original and cleaned light curves, respectively. These tests
434
+ were made following the same procedure as described in Van
435
+ Grootel et al. (2021). The results provided by the inject-and-
436
+ recovery tests show the benefits of using FELIX. In panel 2, we
437
+ might not detect planets with radius ≲1.5 R⊕ for short orbital
438
+ periods ≲3.0 days, and hardly detect planets with ≲2.0–2.5 R⊕
439
+ with orbital periods ≳3.0 days. However, in panel 3, after the
440
+ cleaning performed by FELIX, our detection limits improved. In
441
+ this situation, we started to detect planets with a radius rang-
442
+ ing from 1.0 to 1.5 R⊕ for short orbital periods ≲3.0 days, and
443
+ ∼1.5–2.0 R⊕ for longer orbital periods.
444
+
445
+ We explicitly confirmed, again by injection-and-recovery
446
+ tests, that no transit was removed from the light curve by our
447
+ pre-withening procedure. Transits translate into an LSP by a
448
+ comb of frequencies with decreasing amplitudes, with the orbital
449
+ frequency forb (the highest-amplitude peak) and its harmonics
450
+ n ∗ forb. The number n of harmonics detected depends on the
451
+ transit depth and on the number of transits in the light curve.
452
+ In rare cases, forb might be the only peak detected above our
453
+ usual threshold of 4σ significance. If this orbital frequency is
454
+ of the order of g-mode pulsation frequencies (∼30 min to 3 hr)
455
+ in an identified g-mode pulsator, the forb peak, and hence the
456
+ transit signal, might be removed in our pre-withening procedure
457
+ because it would be misidentified as a g-mode pulsation. This
458
+ situation is quite rare, however, and never occurred during our
459
+ extensive tests.
460
+
461
+ 3 MATRIX
462
+ PlanetHunters/tkmatrix
463
+
464
+ open
465
+
466
+ is
467
+
468
+ A113, page 4 of 12
469
+
470
+ access
471
+
472
+ at
473
+
474
+ https://github.com/
475
+
476
+ 2.2.3. Visual inspection
477
+
478
+ When a SHERLOCK execution was completed, a visual analysis
479
+ of all the signals satisfying the S/N and SDE thresholds was per-
480
+ formed to prove their credibility. Artefacts in the light curve were
481
+ frequently mistaken as transits. This step allowed us to confirm
482
+ or remove this from the list of interesting signals. One exam-
483
+ ple of this visual invalidation is shown in Fig. 3. When a signal
484
+ successfully passed the visual inspection, an analysis of cycle
485
+ 3 data – if available4 – from the same target was carried out
486
+ with the same conditions. The result was then carefully stud-
487
+ ied to detect any signal that matched the one spotted in cycle 1
488
+ (in this phase, signals slightly below thresholds are accounted
489
+ for). If none was detected, the target was set aside and awaited
490
+ potential new observations. If the signal was confirmed, the tar-
491
+ get was worth an individual in-depth investigation and proceeded
492
+ the next steps, which were a literature check. The vetting process
493
+ was triggered as well.
494
+
495
+ 2.2.4. Literature check
496
+
497
+ Databases such as ExoFOP5 and SIMBAD6 were searched for
498
+ a previous detection with similar properties, and we reviewed
499
+ the papers that studied each individual target. If a signal match-
500
+ ing our results was reported previously and a planetary origin
501
+ excluded by the data, the case was closed. This was the case of
502
+ TIC 142875987 (Fig. 1): Bell et al. (2019) determined that this
503
+ transit comes from a low-mass white dwarf closely orbiting the
504
+ hot subdwarf.
505
+
506
+ 2.2.5. Vetting process
507
+
508
+ A vetting procedure of signals satisfying the conditions above
509
+ was made with SHERLOCK using the LATTE package7. For
510
+ each individual transit event, the vetting includes a check of the
511
+ following elements:
512
+
513
+ – The light curve around each transit overplotting TESS’s
514
+
515
+ momentum dumps.
516
+
517
+ – The background flux variation.
518
+ – Monitoring of the x and y positions of the brightest star in
519
+ the aperture as a function of time to test the TESS stability.
520
+
521
+ – The aperture-size dependence.
522
+ – The average flux in and out of the transit.
523
+ – The location and brightness of nearby stars.
524
+ – The light curves from each individual pixel in the target pixel
525
+
526
+ file in transit times.
527
+ More details about LATTE can be found in Eisner et al.
528
+ (2020). Then, we performed a statistical validation using the
529
+ TRICERATOPS package8. TRICERATOPS uses a Bayesian
530
+ framework to compute the probabilities of various astrophysi-
531
+ cal transit-producing scenarios such as a transiting planet with a
532
+ given orbital period (Porb), an eclipsing binary (EB) with Porb,
533
+ or an EB with 2 × Porb (see Table 1 in Giacalone et al. 2021
534
+ for all the scenarios tested). For each scenario, a false positive
535
+
536
+ 4 Only 12 hot subdwarfs out of the 792 were not re-observed during
537
+ cycle 3.
538
+ 5 ExoFOP: Exoplanet Follow-up Observing Program, https://
539
+ exofop.ipac.caltech.edu/tess/
540
+ 6 SIMBAD: Set of Identifications, Measurements and Bibliography for
541
+ Astronomical Data, http://simbad.u-strasbg.fr/simbad/
542
+ 7 LATTE: light curve analysis tool for transiting exoplanet, https:
543
+ //github.com/noraeisner/LATTE
544
+ 8 TRICERATOPS: a tool for vetting and validating TESS objects of
545
+ interest, https://github.com/stevengiacalone/triceratops
546
+
547
+ A. Thuillier et al.: A transit survey to search for planets around hot subdwarfs. II.
548
+
549
+ Fig. 2. sdB hybrid pulsator TIC 169285097 (Gmag = 10.92) observed in Sector 2 by TESS. Panel 1 (top): original (DQUALITY = 0 datapoints)
550
+ PDC-SAP light curve (blue) and light curve after the stellar oscillations were removed (red). Panel 2 (middle): injection-and-recovery test on the
551
+ original light curve. Panel 3 (bottom): injection-and-recovery test on the processed light curve by FELIX. In panels 2 and 3, injected planets have
552
+ radii in the range of 0.5–3.5 R⊕ (steps of 0.37 R⊕) and orbital periods of 0.5–10 days (steps of 0.2 day). For each combination of radius and orbital
553
+ period, we analysed four epochs (or phases), that is, each panel evaluates 760 scenarios computed with our tool MATRIX. Drops in detection at ∼6,
554
+ 8, and 10 days in period are target induced, probably due to noise correlation at these precise periods.
555
+
556
+ probability (FPP) and nearby false positive probability (NFPP)
557
+ are provided, which allow us to verify the reliability of a given
558
+ candidate.
559
+
560
+ 2.2.6. Follow-up of positive signals
561
+
562
+ If a given signal fulfilled the visual inspection, vetting, and statis-
563
+ tical validation steps, an observation with high spatial resolution
564
+ is required to confirm the signal on the target star. This is par-
565
+ ticularly critical for TESS candidates. The pixel scale of TESS
566
+
567
+ is 21′′, and its point-spread function may be as large as 1′, both
568
+ of which increase the probability of contamination by a nearby
569
+ EB (see e.g. Kostov et al. 2019). We did not directly use the
570
+ CROWDNESS factor provided by the TESS team as we indepen-
571
+ dently checked nearby stars at the previous stage of the analysis
572
+ (vetting process). Before triggering a follow-up campaign, we
573
+ first refined the transit parameters. This is because the results
574
+ coming directly from the search step through the modified-
575
+ TLS algorithm are not optimal. The associated uncertainties on
576
+ orbital period (Porb), epoch (T0), and transit duration (d) are too
577
+
578
+ A113, page 5 of 12
579
+
580
+ A&A 664, A113 (2022)
581
+
582
+ parameters and, in particular, its radius. A stellar, white dwarf, or
583
+ brown dwarf origin for the transiting body will need to be ruled
584
+ out based on radial velocity (RV) measurements. We will first
585
+ search for RV data in archives that are open to the community
586
+ (such as the ESO archives) or within the hot subdwarf commu-
587
+ nity. If needed, we write proposals for appropriate spectrographs
588
+ on a case-by-case basis.
589
+
590
+ 3. Results
591
+
592
+ Of the 792 hot subdwarfs identified in the SC observations of
593
+ cycle 1 of TESS, 549 did not have a visually credible transit
594
+ signal. These stars are listed in Table A.1 and their celestial dis-
595
+ tribution are shown in Fig. A.1. Conversely, 243 targets displayed
596
+ at least one signal requiring further investigation. Because some
597
+ targets displayed several interesting signals, their total number
598
+ is greater than the number of targets with positive results: alto-
599
+ gether, we identified 352 potential signals. To keep track of the
600
+ progress of the analysis of this large number of signals, they were
601
+ ranked as follows:
602
+
603
+ – Stage 0: The signal has been detected above our thresholds
604
+
605
+ by SHERLOCK and is visually credible.
606
+
607
+ – Stage 1: A signal with similar properties is detected in the
608
+
609
+ extended mission (cycle 3 of TESS).
610
+
611
+ – Stage 2: The vetting confirms that it does not come from a
612
+ nearby star or from unrelated background features, among
613
+ other checks (see Sect. 2.2.5).
614
+
615
+ – Stage 3: Follow-up observations independently confirm the
616
+
617
+ existence of the signal.
618
+
619
+ – Stage 4: Data confirm the planetary nature of the candidate.
620
+ Of the 352 signals in stage 0, 294 were ruled out (they did not
621
+ match the criterion for stage 1) and 12 belong to stars that were
622
+ not re-observed in cycle 3 (not shown in Table 2). Therefore,
623
+ only 46 went through the vetting and validation process. Thirty
624
+ of them reached stage 2 and became the subject of a follow-
625
+ up campaign currently active (the position in the sky of these
626
+ followed-up targets is plotted in Fig. A.1). At the time of writing,
627
+ 7 targets have been observed and other observations are sched-
628
+ uled for the coming months. No signal of confirmed planetary
629
+ origin has been detected so far. However, several targets have
630
+ already retained our attention, and some of them are detailed in
631
+ the following paragraphs.
632
+
633
+ A first example is the signal detected in the light curve of
634
+ TIC 142875987 (PHL 1539), in which transits were spotted in the
635
+ sector 4 (panels 1–3 of Fig. 1). The transiting body has a period
636
+ of 5.01 days, and the signal is well above our thresholds (S/N
637
+ 30.7, SDE 16.6, see Fig. 1). Therefore data from cycle 3 of this
638
+ target were analysed and the detection was retrieved, as shown
639
+ in panels 4–6 of Fig. 1. However, the literature check revealed
640
+ that this signal has been investigated by Bell et al. (2019), and
641
+ the authors determined that it is a low-mass white dwarf in an
642
+ EB configuration with the targeted sdB star.
643
+
644
+ Another star worth mentioning is TIC 398733009 (KUV
645
+ 02062-0917). A very promising signal was detected in its light
646
+ curve of sector 4, with a period of 0.82 days, and this signal is
647
+ retrieved in sector 30 data (Fig. 4).
648
+
649
+ However, further investigation gave contradictory insights.
650
+ The transits are clear, but
651
+ the eclipses (secondary transits)
652
+ are also detected, which favours the explanation of a self-
653
+ luminous body as a white dwarf. However, the computation
654
+ of the equilibrium temperature of a body that close to its
655
+ host star (∼0.013 AU), following the approach described in
656
+ Charpinet et al. (2011, supplementary information, section D)
657
+
658
+ Fig. 3. Example of a signal in the light curve of TIC 64111698 with
659
+ both S/N and SDE above our thresholds. The light curve was still ruled
660
+ out during our visual inspection. The position of the two detected tran-
661
+ sits is suspicious (top panel): the first transit occurs is in a high-noise
662
+ region, and the second transit is at the edge of a dataset, where typically
663
+ more trends exist. The phase-folded light curve (middle panel) shows
664
+ strong variability out of transit with a similar amplitude as the transit
665
+ signal. Finally, the signal has no harmonics and does not significantly
666
+ rise above the noise (bottom panel). All these hints suggest that the spot-
667
+ ted signal is most likely a false positive.
668
+
669
+ high to schedule an efficient observation. Hence, to refine the
670
+ transit parameters, SHERLOCK used the results coming from
671
+ the modified TLS as priors to perform model fitting, inject-
672
+ ing them into allesfitter (Günther & Daylan 2019, 2021).
673
+ The fitting was made using the dynamic nested sampling algo-
674
+ rithm, whose posterior distributions are much more refined, with
675
+ significant reductions of a few orders of magnitude of the uncer-
676
+ tainties on Porb, T0 , and d. This new set of parameters allowed
677
+ us to schedule a follow-up campaign with reliable observational
678
+ windows.
679
+
680
+ Then, depending on the transit parameters and stellar charac-
681
+ teristics, two options are available to our project. For transit sig-
682
+ nals whose depths exceeded 2.5 mmagand for a stellar brightness
683
+ in the range of 8 to 15 G magnitude, we used the TRAPPIST9
684
+ telescope network. The TRAPPIST network is composed of two
685
+ ground-based 60 cm telescopes, situated in La Silla Observatory
686
+ in Chile (Jehin et al. 2011) and in Oukaïmeden Observatory in
687
+ Morocco (Barkaoui et al. 2019). On the other hand, for signals
688
+ shallower than 2.5 mmag and a stellar brightness ≤14 G magni-
689
+ tude, we used the CHEOPS ESA mission (Benz et al. 2021)10
690
+ via our observational program presented by Van Grootel et al.
691
+ (2021).
692
+ At
693
+
694
+ is successfully recovered by
695
+ the follow-up campaign will trigger an additional investiga-
696
+ tion to characterise the nature of the transiting body. Spectro-
697
+ scopic analysis and spectral energy distribution (SED) fitting
698
+ (Heber et al. 2018) will be carried out to constrain the stellar
699
+
700
+ this step, a signal
701
+
702
+ that
703
+
704
+ 9 TRAPPIST: TRAnsiting Planets and PlanetesImals Small Telescope.
705
+ 10 CHEOPS mission details
706
+ the European
707
+ are
708
+ Space Agency website: https://sci.esa.int/web/cheops/-/
709
+ 54030-summary
710
+
711
+ available
712
+
713
+ on
714
+
715
+ A113, page 6 of 12
716
+
717
+ A. Thuillier et al.: A transit survey to search for planets around hot subdwarfs. II.
718
+
719
+ Table 2. Number of signals at each stage of the analysis.
720
+
721
+ Stage
722
+
723
+ Signals Comment
724
+
725
+ 0
726
+ 1
727
+ 2
728
+ 3
729
+ 4
730
+
731
+ 352
732
+ 46
733
+ 30
734
+ 2
735
+ 0
736
+
737
+ SHERLOCK + visual positive
738
+ Also spotted in cycle 3
739
+ Passed vetting + validation process
740
+ Recovered in follow-up observations
741
+ Planetary nature confirmed
742
+
743
+ Fig. 4. Detection of an interesting transit signal in the light curve of TIC
744
+ 398733009 in sectors 4 (three top panels) and 30 (three bottom panels).
745
+
746
+ gave a result between 5000 and 6000 K in its dayside, assuming
747
+ an albedo of 0.1 and a heat redistribution coefficient β between
748
+ 0.2 and 1 (along with the canonical parameters for the sdB star
749
+ M⋆ = 0.47 M⊙, R⋆ = 0.18 R⊙, and Teff = 31 000 K). This heating
750
+ could explain the secondary transits by itself, as such a hot planet
751
+ emits its own light in the visible part of the spectrum, while the
752
+
753
+ detection of both the transits and eclipses and their regularity
754
+ (as they are separated by half of the orbital period) suggests a
755
+ low eccentricity and an inclination close to 90◦ for the potential
756
+ planet. The vetting phase supported an on-target origin for the
757
+ transits and the absence of contamination from nearby stars, but
758
+ TRICERATOPS gave an abnormally high false-positive prob-
759
+ ability (at 0.99 compared to only 0.45 for a non-false-positive
760
+ probability). We therefore looked for other types of data for this
761
+ target and built a SED from magnitudes in several bands. This
762
+ SED revealed an excess in the red part of the spectrum that is
763
+ too high to be explained by a hot planet. Instead, two stellar
764
+ components are required to explain the observed colours. TIC
765
+ 398733009 is therefore an EB system for which the exact nature
766
+ of the stellar components has to be determined, in particular from
767
+ high-resolution spectroscopy with a high signal-to-noise ratio.
768
+ As this star was confirmed to be no hot subdwarf, it was removed
769
+ from our target list for this project.
770
+
771
+ One last example is the light curve of TIC 369394241
772
+ (HE0452-3654) that was observed in sector 5, in which we
773
+ detected a promising signal that was retrieved in sectors 31 and
774
+ 32. At stage 2 of the analysis (vetting and validation), however,
775
+ we discovered that one of the transits occurred at a time when
776
+ the background flux was very high above the acceptable thresh-
777
+ olds, with a value of several thousands instead of the expected
778
+ few hundreds (Fig. 5). In this case, the procedure was to mask
779
+ the transit and repeat the analysis, in which we did not recover
780
+ the signal. We later discovered that after an update of the SPOC
781
+ algorithms from the TESS team, the portion corresponding to
782
+ this transit is now masked in the most recent version of the PDC-
783
+ SAP light curves. The second transit points are now flagged with
784
+ quality bits 5 (Argabrightening Event) and 12 (Straylight). It was
785
+ therefore labelled as a negative result (no signal) from this point.
786
+
787
+ 4. Discussion
788
+
789
+ We have two main goals in this project. The first goal is to com-
790
+ pute the occurrence rates of planets around hot subdwarfs, and
791
+ the second goal is to provide observational constraints on the sur-
792
+ vival of close-in planets that are engulfed during the RGB phase
793
+ of their host star.
794
+
795
+ Concerning the first goal, we are already in the position to
796
+ draw the very first statistics from the absence of a transit sig-
797
+ nal in 549 of our targets from TESS cycle 1. This non-detection
798
+ scenario only allows us to compute the upper limit of the occur-
799
+ rence rates. We used the same method as in Van Sluijs & Van
800
+ Eylen (2018, Sects. 3.3 and 3.4), itself adapted from Faedi et al.
801
+ (2011), where this limit is written as
802
+
803
+ fmax = 1 − (1 − C)
804
+
805
+ 1
806
+ N′+1 ,
807
+
808
+ (1)
809
+
810
+ with fmax the upper limit for the occurrence rate of planets and
811
+ C the confidence level (between 0 and 1). N′ is defined as the
812
+ product of N, the number of targets in the sample (549) by Ptransit,
813
+ the geometrical transit probability, and Pdetection, the probability
814
+ of detecting a transiting body. The probability Ptransit is directly
815
+ derived from parameters of the system by
816
+
817
+ Ptransit =
818
+
819
+ (cid:33)
820
+
821
+ (cid:32) R⋆ + Rp
822
+ a
823
+
824
+ 1 + e sin(ω)
825
+ 1 − e2
826
+
827
+ ,
828
+
829
+ (2)
830
+
831
+ where R⋆ and Rp are the radii of the star and the planet, a is
832
+ the semi-major axis of the planet, e is the eccentricity, and ω
833
+ is the argument of the periapsis. We assumed circular orbits
834
+ (e = 0), inducing that the right fraction is equal to 1. The only
835
+
836
+ A113, page 7 of 12
837
+
838
+ A&A 664, A113 (2022)
839
+
840
+ Fig. 5. Background flux around the single transits for the detected signal in TIC 369394241 Sector 5. The transits’ position are indicated by orange
841
+ vertical lines. Left and right panels show low and constant values of background flux for the first and third transits. Centre panel shows a steep
842
+ increase of background flux half a day before the second transit, reaching values 20 times higher than the baseline found for the other transits.
843
+
844
+ parameters that matter are therefore Rstar, Rplanet, and a. This
845
+ last parameter is recovered from the period of the planet via
846
+ Kepler’s third law (Kepler 1619). Stellar properties are assumed
847
+ equal to canonical parameters for sdB stars with R⋆ = 0.17 R⊙
848
+ 11
849
+ and M⋆ = 0.47 M⊙. As hot subdwarfs are small stars, we cannot
850
+ neglect planetary radii, hence the equation becomes
851
+
852
+ Ptransit =
853
+
854
+ R⋆ + Rp
855
+ a
856
+
857
+ .
858
+
859
+ (3)
860
+
861
+ The probability Pdetection is computed using injection-and-
862
+ recovery tests for various periods, Rp , and canonical parameters
863
+ for the host stars within pre-determined magnitude bins. Many
864
+ of these injection-and-recovery tests were performed in Van
865
+ Grootel et al. (2021), from which we found that our ability to
866
+ detect planets in light curves provided by the space telescopes
867
+ primarily depends on the magnitude of the star (see their Table 3
868
+ and Fig. 3, in particular).
869
+
870
+ Figure 6 presents the very first estimate for the occurrence
871
+ rates of planets around hot subdwarfs, making the simplification
872
+ that all the 549 stars have G magnitude between 13 and 13.512.
873
+ When our sample is larger and all magnitude bins are more
874
+ populated, occurrence rates will be weighted according to the
875
+ numbers of stars in each magnitude bin. We will also consider
876
+ actual stellar radii.
877
+
878
+ The top panel of Fig. 6 focusses on the injection-and-
879
+ recovery tests, and the bottom panel displays the corresponding
880
+ upper limit fmax for the occurrence of planets for each period-
881
+ radius pair. For example, at a one-day orbital period, we can
882
+ exclude the presence of 3 R⊕ planets in 90% of hot subdwarfs,
883
+ and the presence of 0.5 R⊕ planets in 50% of them. At 6d, these
884
+ numbers decrease to 71% (3 R⊕) and 5% (0.5R⊕). As expected,
885
+ the constraint is less stringent when the detection capabilities are
886
+ lower, which is shown by the higher values for the upper limit for
887
+ smaller planets and longer periods.
888
+
889
+ The second goal of our project is to place observational
890
+ constraints on the survival of close-in planets that are engulfed
891
+
892
+ 11 Canonical radii for hot subdwarf stars are often chosen to be 0.17 and
893
+ 0.18 R⊙ depending on the team or scientist doing the computation, hence
894
+ the difference this value and the value used in Sect. 3. The resulting
895
+ difference is negligible in our case.
896
+ 12 Weaker constraints are expected for fainter stars and stronger con-
897
+ straints for brighter stars. Stars with magnitudes between 13 and 13.5 are
898
+ in the brighter part of our sample (see Van Grootel et al. 2021), but were
899
+ selected for this very first estimate because they are the most numerous
900
+ in the injection-and-recovery tests in our hands, namely, those per-
901
+ formed for Van Grootel et al. (2021). Each computation indeed requires
902
+ thousands of injections and recoveries and can take weeks, therefore we
903
+ relied on those already in our possession.
904
+
905
+ A113, page 8 of 12
906
+
907
+ Fig. 6. Upper limit fmax for the occurrence rate of planets around hot
908
+ subdwarfs from our sample of 549 stars without a detection, considering
909
+ they have a G magnitude between 13 and 13.5. The two panels share
910
+ the same axes with the period (in days) in abscissa and radius of the
911
+ orbiting body (in R⊕) in ordinate. We focus here on short-period planets
912
+ (1 to 6 days) of few Earth radii (0.5–3). Top panel: recovery rate Pdetection
913
+ (deep blue = 0, yellow = 1). Bottom panel: upper limit for the occurrence
914
+ rate fmax (deep blue = 0, yellow = 1). For this graph, the confidence level
915
+ was set to 0.95. For readability, the value of the limit is also written in
916
+ each cell.
917
+
918
+ during the RGB phase of their host star. We wish to list some
919
+ considerations here. Firstly, to achieve this objective, we restrict
920
+
921
+ A. Thuillier et al.: A transit survey to search for planets around hot subdwarfs. II.
922
+
923
+ our sample to sdB stars, which are thought to be in their vast
924
+ majority direct post-RGB stars, the evolutionary paths of sdO
925
+ stars are thought to be more diverse as some may not be direct
926
+ post-RGB stars (see Sect. 1). Secondly, given the nature of the
927
+ transit method, a detection will likely correspond to a close-
928
+ orbiting planet. Three scenarios could explain this configuration:
929
+ it could be (1) a non-engulfed, far-orbiting planet that migrated
930
+ inwards after the end of the RGB phase, (2) a second-generation
931
+ planet formed using remnant material present in the system after
932
+ the ejection of the RGB envelope, and (3) a planet that survived
933
+ the engulfment, at least partially.
934
+
935
+ Although dedicated computations are required, the migration
936
+ of bodies at greater distances that were not engulfed in the enve-
937
+ lope of the red giant star is difficult because sdB stars represent a
938
+ short phase of stellar evolution (∼150 Myr for the core-He burn-
939
+ ing, i.e., EHB, phase; Heber 2016) and planetary migration from
940
+ the outer part of the system is a long process (Mustill et al. 2018).
941
+ For the same reason, the formation of second-generation plan-
942
+ ets is unlikely, in particular in light of the harsh environment
943
+ for planet formation around a hot subdwarf. The survival of an
944
+ engulfed planet is thought to be very difficult, but is completely
945
+ unconstrained from observations. Furthermore, not only does the
946
+ small hot subdwarf size enable the detection of small, possibly
947
+ disintegrating remnant objects, but the ejection of the red giant
948
+ envelope, which is necessary to form a hot subdwarf, may even
949
+ be the reason for the survival of such remnants by stopping the
950
+ spiralling-in inside the host star.
951
+
952
+ Computing the number of expected detections is a difficult
953
+ exercise. Simulations from Staff et al. (2016) indicate a quick
954
+ spiralling of giant planets onto the star. However, some relatively
955
+ low-mass brown dwarfs are known to be very close compan-
956
+ ions to sdB stars (Schaffenroth et al. 2014, 2021) and hence
957
+ have survived this engulfment, so that we can imagine that mas-
958
+ sive planets might survive engulfment in their host star as well.
959
+ Therefore, if any, the most probable bodies able to survive would
960
+ be the cores of former giant planets, whose atmospheres would
961
+ have been stripped out. The exact nature of these cores is thought
962
+ to be a few Earth radii with a dozen Earth masses in the currently
963
+ dominant core-accretion model (Pollack et al. 1996). However,
964
+ we note that the presence of this core is still uncertain even
965
+ for Jupiter (Wahl et al. 2017). For instance, the occurrence rate
966
+ of hot-Jupiter planets around Sun-like stars, which are hot sub-
967
+ dwarf progenitors, is approximately 1% (Wright et al. 2012).
968
+ In the hypothetical case in which all their cores survive, there
969
+ should be 23–50 (for 2300 targets in SC to 5000 targets at all
970
+ cadences; Van Grootel et al. 2021) close-orbiting remnants in
971
+ our sample. Accounting for the transit detection probabilities,
972
+ there should be about two to five detections at orbital periods
973
+ shorter than 10 days in these 2300–5000 targets. However, this
974
+ only includes hot-Jupiter planets. Other types of planets may also
975
+ survive.
976
+
977
+ In an upcoming part of the project, we will also search for
978
+ disintegrating planets, such as debris and/or planets with dust
979
+ tails, with a new module that we are building in the SHERLOCK
980
+ pipeline. This upgrade will allow us to search for transits with
981
+ asymmetric shapes that are produced by the dust tails, such as
982
+ the one found around a white dwarf by Vanderburg et al. (2015).
983
+ One last point is the potential role that giant planets might have in
984
+ the formation of hot subdwarfs in the case of an engulfment. If it
985
+ is large enough, the planet might help to expel the envelope, but
986
+ this possibility is debated (e.g. Kramer et al. 2020). The detection
987
+ of remnants around apparently single hot subdwarfs may provide
988
+ new elements to settle this issue.
989
+
990
+ Acknowledgements. We warmly thank the anonymous referee for constructive
991
+ remarks that improved our paper. We thank Uli Heber and Elizabeth M. Green
992
+ for their help on the characterisation of several of our targets, as well as atten-
993
+ dees of the sdOB9.5 conference in Potsdam, namely but not limited to, Stephan
994
+ Geier and Philipp Podsiadlowksi, as the discussion there were of great interest for
995
+ this work. This work has been supported by the University of Liège through an
996
+ ARC grant for Concerted Research Actions financed by the Wallonia-Brussels
997
+ Federation. A.T. acknowledge financial support from the ULB “Fond de rat-
998
+ trapage PDR”. V.V.G. and L.S are senior F.R.S.-FNRS Research Associates.
999
+ S.C. acknowledges financial support from the Centre National d’Études Spatiales
1000
+ (CNES, France). This paper includes data collected by the TESS mission. Fund-
1001
+ ing for the TESS mission is provided by the NASA Explorer Program. Funding
1002
+ for the TESS Asteroseismic Science Operations Centre is provided by the Danish
1003
+ National Research Foundation (Grant agreement no.: DNRF106), ESA PRODEX
1004
+ (PEA 4000119301) and Stellar Astrophysics Centre (SAC) at Aarhus Univer-
1005
+ sity. We thank the TESS team and staff and TASC/TASOC for their support
1006
+ of the present work. This work has made use of data from the ESA mission Gaia
1007
+ (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Process-
1008
+ ing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/
1009
+ gaia/dpac/consortium). Funding for the DPAC has been provided by national
1010
+ institutions, in particular the institutions participating in the Gaia Multilateral
1011
+ Agreement.
1012
+
1013
+ References
1014
+
1015
+ Barkaoui, K., Burdanov, A., Hellier, C., et al. 2019, AJ, 157, 43
1016
+ Bear, E., & Soker, N. 2014, MNRAS, 437, 1400
1017
+ Bell, K. J., Kosakowski, A., Kilic, M., et al. 2019, RNAAS, 3, 81
1018
+ Benz, W., Broeg, C., Fortier, A., et al. 2021, Exp. Astron., 51, 109
1019
+ Blokesz, A., Krzesinski, J., & Kedziora-Chudczer, L. 2019, A&A, 627, A86
1020
+ Borucki, W. J., Koch, D., Basri, G., et al. 2010, Science, 327, 977
1021
+ Charpinet, S., Green, E. M., Baglin, A., et al. 2010, A&A, 516, A6
1022
+ Charpinet, S., Fontaine, G., Brassard, P., et al. 2011, Nature, 480, 496
1023
+ Deeming, T. J. 1975, Ap&SS, 36, 137
1024
+ Dorman, B., Rood, R. T., & O’Connell, R. W. 1993, ApJ, 419, 596
1025
+ Dévora-Pajares, M., & Pozuelos, F. J. 2022, MATRIX: Multi-phAse Transits
1026
+
1027
+ Recovery from Injected eXoplanets
1028
+
1029
+ Eisner, N., Lintott, C., & Aigrain, S. 2020, J. Open Source Softw., 5, 2101
1030
+ Faedi, F., West, R. G., Burleigh, M. R., Goad, M. R., & Hebb, L. 2011, MNRAS,
1031
+
1032
+ 410, 899
1033
+
1034
+ Giacalone, S., Dressing, C. D., Jensen, E. L. N., et al. 2021, AJ, 161, 24
1035
+ Green, E. M., Fontaine, G., Hyde, E. A., For, B. Q., & Chayer, P. 2008, in
1036
+ Astronomical Society of the Pacific Conference Series, Hot Subdwarf Stars
1037
+ and Related Objects, eds. U. Heber, C. S. Jeffery, & R. Napiwotzki, 392,
1038
+ 75
1039
+
1040
+ Günther, M. N., & Daylan, T. 2019, Allesfitter: Flexible Star and Exoplanet
1041
+ Inference From Photometry and Radial Velocity, Astrophysics Source Code
1042
+ Library [record ascl:1903.003]
1043
+
1044
+ Günther, M. N., & Daylan, T. 2021, ApJS, 254, 13
1045
+ Heber, U. 1986, A&A, 155, 33
1046
+ Heber, U. 2016, PASP, 128, 082001
1047
+ Heber, U., Irrgang, A., & Schaffenroth, J. 2018, Open Astron., 27, 35
1048
+ Hippke, M., & Heller, R. 2019, A&A, 623, A39
1049
+ Hippke, M., David, T. J., Mulders, G. D., & Heller, R. 2019, AJ, 158, 143
1050
+ Howell, S. B., Sobeck, C., Haas, M., et al. 2014, PASP, 126, 398
1051
+ Iben, I., J. 1990, ApJ, 353, 215
1052
+ Iben, I. J., & Tutukov, A. V. 1984, ApJS, 54, 335
1053
+ Jehin, E., Gillon, M., Queloz, D., et al. 2011, The Messenger, 145, 2
1054
+ Johnson, C., Green, E., Wallace, S., et al. 2014, in Astronomical Society of the
1055
+ Pacific Conference Series, 6th Meeting on Hot Subdwarf Stars and Related
1056
+ Objects, eds. V. van Grootel, E. Green, G. Fontaine, & S. Charpinet, 481,
1057
+ 153
1058
+
1059
+ Jones, M. I., & Jenkins, J. S. 2014, A&A, 562, A129
1060
+ Kepler, J. 1619, Ioannis Keppleri harmonices mundi libri V (Austria: Linz)
1061
+ Kostov, V. B., Schlieder, J. E., Barclay, T., et al. 2019, AJ, 158, 32
1062
+ Kramer, M., Schneider, F. R. N., Ohlmann, S. T., et al. 2020, A&A, 642,
1063
+
1064
+ A97
1065
+
1066
+ Krzesinski, J. 2015, A&A, 581, A7
1067
+ Krzesinski, J., Blokesz, A., Ogłoza, W., & Dró˙zd˙z, M. 2020, in Origins: From
1068
+ the Protosun to the First Steps of Life, eds. B. G. Elmegreen, L. V. Tóth, &
1069
+ M. Güdel, 345, 306
1070
+
1071
+ Miller Bertolami, M. M., Althaus, L. G., Unglaub, K., & Weiss, A. 2008, A&A,
1072
+
1073
+ 491, 253
1074
+
1075
+ Mustill, A. J., Villaver, E., Veras, D., Gänsicke, B. T., & Bonsor, A. 2018,
1076
+
1077
+ MNRAS, 476, 3939
1078
+
1079
+ A113, page 9 of 12
1080
+
1081
+ A&A 664, A113 (2022)
1082
+
1083
+ Oreiro Rey, R., Rodríguez López, C., Ulla, A., et al. 2004, in Lecture Notes and
1084
+
1085
+ Essays in Astrophysics, eds. A. Ulla, & M. Manteiga, 1, 133
1086
+
1087
+ Pollack, J. B., Hubickyj, O., Bodenheimer, P., et al. 1996, Icarus, 124, 62
1088
+ Pozuelos, F. J., Suárez, J. C., de Elía, G. C., et al. 2020, A&A, 641, A23
1089
+ Reindl, N., Geier, S., Kupfer, T., et al. 2016, A&A, 587, A101
1090
+ Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2014, in Society of Photo-
1091
+ Optical Instrumentation Engineers (SPIE) Conference Series, Space Tele-
1092
+ scopes and Instrumentation 2014: Optical, Infrared, and Millimeter Wave,
1093
+ 9143, 914320
1094
+
1095
+ Saffer, R. A., Bergeron, P., Koester, D., & Liebert, J. 1994, ApJ, 432, 351
1096
+ Saio, H., & Jeffery, C. S. 2000, MNRAS, 313, 671
1097
+ Saio, H., & Jeffery, C. S. 2002, MNRAS, 333, 121
1098
+ Schaffenroth, V., Geier, S., Heber, U., et al. 2014, A&A, 564, A98
1099
+ Schaffenroth, V., Casewell, S. L., Schneider, D., et al. 2021, MNRAS, 501,
1100
+
1101
+ 3847
1102
+
1103
+ Setiawan, J., Klement, R. J., Henning, T., et al. 2010, Science, 330, 1642
1104
+
1105
+ Silvotti, R., Schuh, S., Janulis, R., et al. 2007, Nature, 449, 189
1106
+ Silvotti, R., Schuh, S., Kim, S. L., et al. 2018, A&A, 611, A85
1107
+ Silvotti, R., Ostensen, R. H., & Telting, J. H. 2020 https://doi.org/10.
1108
+
1109
+ 5281/zenodo.3588477
1110
+
1111
+ Staff, J. E., De Marco, O., Wood, P., Galaviz, P., & Passy, J.-C. 2016, MNRAS,
1112
+
1113
+ 458, 832
1114
+
1115
+ Van Grootel, V., Pozuelos, F. J., Thuillier, A., et al. 2021, A&A, 650,
1116
+
1117
+ A205
1118
+
1119
+ Vanderburg, A., Johnson, J. A., Rappaport, S., et al. 2015, Nature, 526, 546
1120
+ Van Sluijs, L., & Van Eylen, V. 2018, MNRAS, 474, 4603
1121
+ Villaver, E., & Livio, M. 2009, ApJ, 705, L81
1122
+ Wahl, S. M., Hubbard, W. B., Militzer, B., et al. 2017, Geophys. Res. Lett., 44,
1123
+
1124
+ 4649
1125
+
1126
+ Webbink, R. F. 1984, ApJ, 277, 355
1127
+ Wright, J. T., Marcy, G. W., Howard, A. W., et al. 2012, ApJ, 753, 160
1128
+ Zong, W., Charpinet, S., & Vauclair, G. 2016, A&A, 594, A46
1129
+
1130
+ A113, page 10 of 12
1131
+
1132
+ A. Thuillier et al.: A transit survey to search for planets around hot subdwarfs. II.
1133
+
1134
+ Appendix A: List of stars without planetary transit
1135
+
1136
+ signals above thresholds
1137
+
1138
+ Table A.1. List of targets (ranked by increasing TIC number) without a detected planetary transit signal above our thresholds (they
1139
+ did not reach stage 1 of the analysis). See details in Sect. 3.
1140
+
1141
+ 8842
1142
+ 1526470
1143
+ 2213137
1144
+ 3990402
1145
+ 5051080
1146
+ 6593243
1147
+ 7319744
1148
+ 9053429
1149
+ 9102069
1150
+ 9268360
1151
+ 9358354
1152
+ 10932480
1153
+ 12378718
1154
+ 12528447
1155
+ 12549312
1156
+ 13069774
1157
+ 13145616
1158
+ 20448010
1159
+ 24448566
1160
+ 24908871
1161
+ 25136499
1162
+ 25137423
1163
+ 25245570
1164
+ 25300088
1165
+ 25948892
1166
+ 28052378
1167
+ 29821374
1168
+ 29840077
1169
+ 30019744
1170
+ 30415110
1171
+ 31174073
1172
+ 32354769
1173
+ 32661254
1174
+ 33318760
1175
+ 33321190
1176
+ 33490778
1177
+ 33526769
1178
+ 33767134
1179
+ 33944704
1180
+ 36729387
1181
+ 36995993
1182
+ 37602756
1183
+ 38423413
1184
+ 38511369
1185
+ 44625421
1186
+ 45362220
1187
+ 47352092
1188
+ 47355413
1189
+ 47377536
1190
+ 47482655
1191
+ 49711130
1192
+ 50384080
1193
+ 53939726
1194
+ 54986420
1195
+ 56124677
1196
+ 56648314
1197
+ 59769766
1198
+ 61029108
1199
+
1200
+ 61379646
1201
+ 61621625
1202
+ 61728030
1203
+ 62023018
1204
+ 62223401
1205
+ 62381958
1206
+ 62483415
1207
+ 63428034
1208
+ 63695388
1209
+ 63696810
1210
+ 64008380
1211
+ 64111698
1212
+ 64112207
1213
+ 65145453
1214
+ 65324889
1215
+ 66398320
1216
+ 66493797
1217
+ 66626624
1218
+ 67583991
1219
+ 67584818
1220
+ 67598107
1221
+ 69841801
1222
+ 69911593
1223
+ 70723238
1224
+ 70776745
1225
+ 70962870
1226
+ 71013467
1227
+ 71109189
1228
+ 71133157
1229
+ 71150825
1230
+ 71248239
1231
+ 71345281
1232
+ 71410075
1233
+ 71716888
1234
+ 72667488
1235
+ 72763826
1236
+ 73185296
1237
+ 73764693
1238
+ 76184341
1239
+ 77005197
1240
+ 77360048
1241
+ 77477081
1242
+ 77959296
1243
+ 77992461
1244
+ 79246796
1245
+ 79493362
1246
+ 79956635
1247
+ 79958453
1248
+ 80057233
1249
+ 80170223
1250
+ 80290366
1251
+ 80427831
1252
+ 83755080
1253
+ 88417452
1254
+ 89148712
1255
+ 89467049
1256
+ 89529774
1257
+ 91316983
1258
+
1259
+ 91556284
1260
+ 92226759
1261
+ 92984690
1262
+ 94945019
1263
+ 98868202
1264
+ 98871628
1265
+ 101403951
1266
+ 101454745
1267
+ 101512805
1268
+ 101540375
1269
+ 101545865
1270
+ 101817287
1271
+ 101817673
1272
+ 107548305
1273
+ 111184069
1274
+ 113118441
1275
+ 114115352
1276
+ 114502553
1277
+ 114807149
1278
+ 115175941
1279
+ 115273584
1280
+ 115692439
1281
+ 117626475
1282
+ 117735611
1283
+ 118327563
1284
+ 118412596
1285
+ 118588732
1286
+ 120580052
1287
+ 120610731
1288
+ 121318590
1289
+ 121550523
1290
+ 122521574
1291
+ 123665218
1292
+ 124029363
1293
+ 124175842
1294
+ 125556577
1295
+ 126637970
1296
+ 126803779
1297
+ 128971400
1298
+ 131343095
1299
+ 138707463
1300
+ 138707823
1301
+ 139208432
1302
+ 139266474
1303
+ 139682931
1304
+ 139804925
1305
+ 140494440
1306
+ 142200764
1307
+ 142875987
1308
+ 143058705
1309
+ 143699381
1310
+ 143923307
1311
+ 144193687
1312
+ 144306296
1313
+ 144804862
1314
+ 146251617
1315
+ 146282446
1316
+ 146323153
1317
+
1318
+ 146437397
1319
+ 146520657
1320
+ 147115112
1321
+ 147283842
1322
+ 147349694
1323
+ 148993614
1324
+ 149767908
1325
+ 149846788
1326
+ 151892844
1327
+ 152286180
1328
+ 152373379
1329
+ 152374958
1330
+ 153165824
1331
+ 155776755
1332
+ 156215534
1333
+ 157323544
1334
+ 158235404
1335
+ 158335560
1336
+ 158343496
1337
+ 159640970
1338
+ 159669717
1339
+ 159693368
1340
+ 159747831
1341
+ 159805154
1342
+ 159833896
1343
+ 159843438
1344
+ 160078278
1345
+ 160797304
1346
+ 161153327
1347
+ 161402643
1348
+ 164754858
1349
+ 165110581
1350
+ 166395799
1351
+ 166748749
1352
+ 167456967
1353
+ 167746025
1354
+ 167976324
1355
+ 168304840
1356
+ 169285097
1357
+ 170108378
1358
+ 170203297
1359
+ 170869314
1360
+ 176380024
1361
+ 177079697
1362
+ 178345419
1363
+ 178873605
1364
+ 178893906
1365
+ 178896096
1366
+ 179034615
1367
+ 179278778
1368
+ 181820016
1369
+ 181914779
1370
+ 182012072
1371
+ 183530773
1372
+ 183563651
1373
+ 186149538
1374
+ 188174892
1375
+ 189585096
1376
+
1377
+ 189585273
1378
+ 193087619
1379
+ 193092806
1380
+ 197570382
1381
+ 197614201
1382
+ 197687846
1383
+ 197693940
1384
+ 197765610
1385
+ 200388625
1386
+ 200436339
1387
+ 200545781
1388
+ 201690280
1389
+ 206347723
1390
+ 206462125
1391
+ 206469610
1392
+ 206482194
1393
+ 206590489
1394
+ 207208668
1395
+ 207328336
1396
+ 209393544
1397
+ 209397773
1398
+ 212320065
1399
+ 212347311
1400
+ 212352349
1401
+ 214568914
1402
+ 218960493
1403
+ 218961596
1404
+ 219225205
1405
+ 219974863
1406
+ 220026025
1407
+ 220347928
1408
+ 220370211
1409
+ 220472655
1410
+ 220476769
1411
+ 220479184
1412
+ 220573709
1413
+ 228508601
1414
+ 229050493
1415
+ 229051528
1416
+ 229144939
1417
+ 231308507
1418
+ 231629787
1419
+ 231695087
1420
+ 231712886
1421
+ 231812407
1422
+ 234281664
1423
+ 234295068
1424
+ 234386436
1425
+ 234526947
1426
+ 235011371
1427
+ 237322080
1428
+ 237338096
1429
+ 238853890
1430
+ 241675531
1431
+ 246881248
1432
+ 246881770
1433
+ 248382748
1434
+ 248391428
1435
+
1436
+ 248949857
1437
+ 253656464
1438
+ 253932935
1439
+ 254263628
1440
+ 254287117
1441
+ 255697796
1442
+ 259392878
1443
+ 259539930
1444
+ 259864042
1445
+ 260749705
1446
+ 260795163
1447
+ 260839766
1448
+ 261241692
1449
+ 261262954
1450
+ 261380566
1451
+ 261427146
1452
+ 261679852
1453
+ 262039948
1454
+ 262818577
1455
+ 262921847
1456
+ 262927968
1457
+ 265445890
1458
+ 266680031
1459
+ 269762836
1460
+ 269855226
1461
+ 269973828
1462
+ 270000741
1463
+ 270071350
1464
+ 270245694
1465
+ 270285517
1466
+ 270380474
1467
+ 270394275
1468
+ 270416019
1469
+ 270562073
1470
+ 271576664
1471
+ 273218137
1472
+ 273862178
1473
+ 274035031
1474
+ 275185985
1475
+ 275921038
1476
+ 277773221
1477
+ 277892210
1478
+ 278659026
1479
+ 278705842
1480
+ 279342801
1481
+ 280051980
1482
+ 280775141
1483
+ 280789378
1484
+ 281595262
1485
+ 281658096
1486
+ 281851153
1487
+ 283866221
1488
+ 284677903
1489
+ 284703017
1490
+ 286099192
1491
+ 289533626
1492
+ 289721074
1493
+ 289737935
1494
+
1495
+ 289795477
1496
+ 289823802
1497
+ 290401825
1498
+ 290646079
1499
+ 293165262
1500
+ 293463617
1501
+ 293520466
1502
+ 293791715
1503
+ 294836239
1504
+ 300607115
1505
+ 301022374
1506
+ 301405970
1507
+ 301454053
1508
+ 304103779
1509
+ 307934757
1510
+ 309754884
1511
+ 317058444
1512
+ 317154662
1513
+ 317350896
1514
+ 317439554
1515
+ 319602897
1516
+ 320092330
1517
+ 320173712
1518
+ 320176500
1519
+ 320417198
1520
+ 320484212
1521
+ 320529836
1522
+ 320533176
1523
+ 320591159
1524
+ 320591947
1525
+ 320660807
1526
+ 320939631
1527
+ 320951753
1528
+ 320965274
1529
+ 322285377
1530
+ 325394866
1531
+ 325566833
1532
+ 326096162
1533
+ 326328277
1534
+ 326453105
1535
+ 328179553
1536
+ 332697630
1537
+ 332701732
1538
+ 332742020
1539
+ 333659638
1540
+ 339262037
1541
+ 339381798
1542
+ 339907982
1543
+ 341001113
1544
+ 343828974
1545
+ 345449417
1546
+ 347412256
1547
+ 349367583
1548
+ 350155206
1549
+ 350273432
1550
+ 350583903
1551
+ 352142391
1552
+ 352315023
1553
+
1554
+ 352480413
1555
+ 354572272
1556
+ 355002205
1557
+ 360736086
1558
+ 360804908
1559
+ 360806208
1560
+ 364424541
1561
+ 369377398
1562
+ 369965957
1563
+ 371801053
1564
+ 372172495
1565
+ 373245897
1566
+ 374408958
1567
+ 379352979
1568
+ 380274928
1569
+ 380590152
1570
+ 382247276
1571
+ 382375326
1572
+ 382383606
1573
+ 382415345
1574
+ 382416576
1575
+ 383217734
1576
+ 386642921
1577
+ 386653873
1578
+ 386703105
1579
+ 387107334
1580
+ 389476185
1581
+ 389520459
1582
+ 389752750
1583
+ 392703299
1584
+ 392758248
1585
+ 393033236
1586
+ 393491149
1587
+ 393941149
1588
+ 394018151
1589
+ 394494920
1590
+ 394496912
1591
+ 394517648
1592
+ 394517681
1593
+ 394678374
1594
+ 394698511
1595
+ 396004353
1596
+ 396695965
1597
+ 398796167
1598
+ 398801916
1599
+ 403317759
1600
+ 404158419
1601
+ 404430996
1602
+ 404467239
1603
+ 405799245
1604
+ 406239686
1605
+ 406241063
1606
+ 406242857
1607
+ 406280054
1608
+ 406280906
1609
+ 407768021
1610
+ 409644971
1611
+ 409850857
1612
+
1613
+ 410135274
1614
+ 410280069
1615
+ 410390905
1616
+ 413300076
1617
+ 415339307
1618
+ 419571221
1619
+ 419998091
1620
+ 420049852
1621
+ 421895532
1622
+ 421939894
1623
+ 421951567
1624
+ 421999342
1625
+ 422004014
1626
+ 422043417
1627
+ 422149668
1628
+ 423160573
1629
+ 423761655
1630
+ 423763742
1631
+ 424881855
1632
+ 424941595
1633
+ 425064757
1634
+ 425206178
1635
+ 425516683
1636
+ 425799931
1637
+ 426030266
1638
+ 432162081
1639
+ 432744391
1640
+ 436579904
1641
+ 436636292
1642
+ 436682542
1643
+ 437237493
1644
+ 439461184
1645
+ 439905042
1646
+ 441399312
1647
+ 441401311
1648
+ 441499030
1649
+ 441508514
1650
+ 443625969
1651
+ 445927286
1652
+ 455206965
1653
+ 457168745
1654
+ 457196370
1655
+ 459932445
1656
+ 464549287
1657
+ 464769756
1658
+ 466229760
1659
+ 466273074
1660
+ 466310972
1661
+ 469791892
1662
+ 469981019
1663
+ 471015202
1664
+ 471015203
1665
+ 471015416
1666
+
1667
+ A113, page 11 of 12
1668
+
1669
+ A&A 664, A113 (2022)
1670
+
1671
+ Fig. A.1. Celestial distribution of the targets without a detected signal (blue crosses) and those that reached stage 3 of the analysis
1672
+ and for which follow up is scheduled (open purple circles). See also Fig. 1 from Van Grootel et al. (2021) for the global distribution
1673
+ of hot subdwarf stars observed by the missions Kepler, K2, TESS, and CHEOPS. Data shown in equatorial coordinates and in a
1674
+ Mollweide projection of the sky.
1675
+
1676
+ A113, page 12 of 12
1677
+
1678
+
documents/txt/armstrong2020.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:334ebc8e1f280511d5fb0570cb7f9d766a63ccc67b6374e98b1be64ec65cf0ea
3
+ size 5983166
documents/txt/author.txt ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ The author of the Phd. Thesis is Martín Dévora Pajares. He is a senior software engineer, data scientist and a Phd.
2
+ student on the Granada University, Spain. His supervisors are Francisco J. Pozuelos Romero, Senior Post-doc researcher at the
3
+ Andalusian Astrophysics Institute (iaa.es) and Juan Carlos Suárez Yanes, Senior Phd. Professor at the Granada
4
+ University (ugr.es). Martín has developed SHERLOCK, WATSON and MATRIX and is also actively maintaining them together to the development
5
+ of new tools and neural networks.
documents/txt/dvr1.txt ADDED
The diff for this file is too large to render. See raw diff
 
documents/txt/dvr2.txt ADDED
@@ -0,0 +1,4273 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2
+ © 2019. The Astronomical Society of the Pacific. Printed in the U.S.A.
3
+
4
+ https://doi.org/10.1088/1538-3873/aaf44d
5
+
6
+ Kepler Data Validation II–Transit Model Fitting and Multiple-planet Search
7
+
8
+ Jie Li1, Peter Tenenbaum1, Joseph D. Twicken1
9
+
10
+ , Christopher J. Burke1, Jon M. Jenkins2, Elisa V. Quintana3,
11
+
12
+ Jason F. Rowe1, and Shawn E. Seader4
13
+ 1 SETI Institute, Mountain View, CA, 94043, USA; jie.li-1@nasa.gov
14
+ 2 NASA Ames Research Center, Moffett Field, CA, 94035, USA
15
+ 3 NASA Goddard Space Flight Center, Greenbelt, MD, 20771, USA
16
+ 4 Rincon Research Corporation, Tucson, AZ, 85711, USA
17
+ Received 2018 October 22; accepted 2018 November 26; published 2019 January 14
18
+
19
+ Abstract
20
+ This paper discusses the transit model-fitting and multiple-planet search algorithms and performance of the Kepler
21
+ Science Data Processing Pipeline, developed by the Kepler Science Operations Center (SOC). Threshold crossing
22
+ events (TCEs), which are transit candidate events, are generated by the Transiting Planet Search (TPS) component
23
+ of the pipeline and subsequently processed in the data validation (DV) component. The transit model is used in DV
24
+ to fit TCEs to characterize planetary candidates and to derive parameters that are used in various diagnostic tests to
25
+ classify them. After the signature associated with the TCE is removed from the light curve of the target star, the
26
+ residual light curve goes through TPS again to search for additional TCEs. The iterative process of transit model-
27
+ fitting and multiple-planet search continues until no TCE is generated from the residual light curve or an upper
28
+ limit is reached. The transit model-fitting and multiple-planet search performance of the final release (9.3, 2016
29
+ January) of the pipeline is demonstrated with the results of the processing of four years (17 quarters) of flight data
30
+ from the primary Kepler Mission. The transit model-fitting results are accessible from the NASA Exoplanet
31
+ Archive. The final version of the SOC codebase is available through GitHub.
32
+ Key words: methods: data analysis – methods: statistical – techniques: image processing –
33
+ techniques: photometric – planets and satellites: detection
34
+
35
+ 1. Introduction
36
+ This paper discusses transit model-fitting and multiple-planet
37
+ search algorithms and performance that are part of the Data
38
+ Validation (DV) component of
39
+ the Kepler Science Data
40
+ Processing Pipeline (Jenkins et al. 2010a), developed by the
41
+ Kepler Science Operations Center (SOC) at NASA Ames
42
+ Research Center. An introduction to Kepler Mission, the Kepler
43
+ Science Data Processing Pipeline and the DV component is
44
+ provided in a companion paper (Twicken et al. 2018), which
45
+ also details the DV diagnostic tests and data products for vetting
46
+ transiting planet candidates.
47
+
48
+ The transit model fitting is designed for the following three
49
+ main tasks: (1) the orbital property and the nature of the planetary
50
+ candidates are characterized; (2) the fitted parameters of the transit
51
+ model and the corresponding light curve generated from the
52
+ model are used in the diagnostic tests in DV to aid in the
53
+ assessment and classification of planetary candidates; (3) when
54
+ the Transiting Planet Search (TPS) component is called, only one
55
+ Threshold Crossing Event (TCE) with the maximum multiple
56
+
57
+ Original content from this work may be used under the terms
58
+ of the Creative Commons Attribution 3.0 licence. Any further
59
+ distribution of this work must maintain attribution to the author(s) and the title
60
+ of the work, journal citation and DOI.
61
+
62
+ event detection statistic (MES) is generated. To search for
63
+ transit
64
+ multiple-planetary candidates, an iterative process of
65
+ model-fitting and multiple-planet search is implemented in DV.
66
+ For each target star, the transit model parameters are fitted to each
67
+ TCE generated by TPS, the signature of known TCEs is removed
68
+ from the light curve, and then the residual is provided to TPS
69
+ again to search for additional TCEs. This iteration will only
70
+ terminate once no new TCEs are identified or a preset upper limit
71
+ is reached (set to 10 for the SOC 9.3 run producing the Data
72
+ Release (DR) 25 TCEs).
73
+
74
+ The transit model-fitting results, such as the fitted parameters
75
+ and uncertainties; derived parameters and uncertainties; good-
76
+ ness-of-fit metrics; and the diagnostic plots, are included
77
+ in comprehensive DV reports by target, and one-page DV
78
+ summary reports by TCE. The reports and summaries are
79
+ accessible by the science community at the Exoplanet Archive5
80
+ at the NASA Exoplanet Science Institute (NExScI; Akeson
81
+ et al. 2013). The final version of the SOC 9.3 codebase is
82
+ available to the general public through GitHub.6
83
+
84
+ The transit model-fitting and multiple-planet search algo-
85
+ rithm in the initial revision of DV (SOC 6.1) was described by
86
+
87
+ 5 https://exoplanetarchive.ipac.caltech.edu
88
+ 6 https://github.com/nasa/kepler-pipeline
89
+
90
+ 1
91
+
92
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
93
+
94
+ Li et al.
95
+
96
+ Tenenbaum et al. (2010) and Wu et al. (2010). DV evolved
97
+ greatly since then. Major changes in the transit model-fitting
98
+ and multiple-planet search algorithm include (1) the transit
99
+ model described in Tenenbaum et al. (2010) and Wu et al.
100
+ (2010) is changed to the geometric transit model, including a
101
+ nonlinear limb-darkening model (Claret & Bloemen 2011), for
102
+ a better modeling accuracy; (2) the reduced-parameter fits are
103
+ added; and (3) the trapezoidal model fit is added.
104
+
105
+ Iterative transit fitting and multiple-planet search has been
106
+ done extensively by various groups. Foreman-Mackey et al.
107
+ (2015) performs a joint fit of the transit model and systematics,
108
+ which may be more sensitive than the algorithm used in the
109
+ SOC 9.3 codebase but is computationally more expensive.
110
+ Crossfield et al. (2016), Crossfield et al. (2018), Petigura et al.
111
+ (2018), and Yu et al. (2018) use the “TERRA” software
112
+ package and presume the systematic error correction has
113
+ whitened the colored noise of the light curve. Dressing &
114
+ Charboneau (2015), Vanderburg et al. (2016), and Rizzuto
115
+ et al. (2017) use box least-squares algorithm and also assume
116
+ that the residual observation noise is white. In this paper, the
117
+ transit model fitting is implemented with an iterative loop that
118
+ includes a whitening filter and a transit fitter. In addition,
119
+ compared to the similar work by other groups, the reduced-
120
+ parameter fits described in this paper improve the consistency
121
+ of the results of the geometric transit model fit, and the
122
+ trapezoidal model fit provides a quick assessment of the transit
123
+ signal.
124
+
125
+ In this paper, the final SOC 9.3 codebase is described. The
126
+ architecture of transit model-fitting and multiple-planet search
127
+ algorithm is described in Section 2, and the light curve
128
+ preprocessing procedures are described in Section 3. The
129
+ geometric transit model is described in Sections 4. Section 5
130
+ describes how a synthetic light curve is generated from the
131
+ fitted parameters of the geometric transit model, and Section 6
132
+ describes the algorithms to fit
133
+ the light curves with the
134
+ geometric transit model. A fitting algorithm with the trapezoi-
135
+ dal model is described in Section 7, and the multiple-planet
136
+ search is discussed in Section 8. The performance of the transit
137
+ model-fitting and multiple-planet search is demonstrated in
138
+ Section 9. Finally, conclusions are presented in Section 10.
139
+
140
+ 2. Architecture of Transit Model Fitting and
141
+ Multiple-planet Search
142
+ This section describes the architecture of transit model-fitting
143
+ shown in the
144
+ search algorithm. As
145
+
146
+ and multiple-planet
147
+ flowchart in Figure 1, it is an iterative process.
148
+
149
+ the
150
+ When a TCE is generated by the TPS component,
151
+ corresponding systematic error-corrected light curve of the
152
+ target star, generated by the pre-search data conditioning (PDC)
153
+ component of the pipeline, is furnished to DV along with the
154
+ including the
155
+ transit parameters associated with the TCE,
156
+
157
+ transit epoch (central time of first transit), orbital period, transit
158
+ duration, and MES of the TCE. The light curve may span one
159
+ or more observing quarters. After
160
+ several preprocessing
161
+ procedures, the light curve of the target star goes through a
162
+ transit model-fitting algorithms, which include
163
+ series of
164
+ reduced-parameter fits, all-transit fit, odd–even transit fit and
165
+ trapezoidal model fit.
166
+
167
+ As shown in Figure 1, the preprocessed light curve is first
168
+ subjected to a set of reduced-parameter fits, in which the impact
169
+ parameter is set to fixed values of 0.1, 0.3, 0.5, 0.7, and 0.9,
170
+ and only the parameters of transit epoch time, planet orbital
171
+ period, ratio of planet radius to star radius, and ratio of
172
+ semimajor axis to star radius are fitted to a geometric transit
173
+ model. The initial values of the fitted parameters of the
174
+ reduced-parameter fits are determined from the TCE para-
175
+ meters. The reduced-parameter fits resolve the degenerate
176
+ problem of fitting the impact parameter, which is discussed in
177
+ Section 6.2. After the completion of the reduced-parameter fits,
178
+ all-transit fit and odd–even transit fit follow, in which the fitting
179
+ algorithms are applied to all transits, odd transits and even
180
+ transits, respectively. The all-transit fit and odd–even transit fit
181
+ are both initialized with the fitted parameters of the reduced-
182
+ parameter fit with the minimum χ2 metric. The output of the
183
+ all-transit fit is used in several diagnostic tests of DV and the
184
+ assessment of planet candidacy, and the output of the odd–even
185
+ transit fit is used in a specific DV diagnostic test to identify
186
+ false positives due to an eclipsing binary target or a target with
187
+ an eclipsing binary in the background. In addition to the fitting
188
+ algorithms with the geometric transit model, a fitting algorithm
189
+ is implemented. As shown in
190
+ with the trapezoidal model
191
+ Figure 1, an alternative detrending and normalization algorithm
192
+ is applied to the PDC light curve prior to the trapezoidal model
193
+ fit. The output of the trapezoidal model ��t is used in the
194
+ diagnostic tests of DV when the fit with the geometric transit
195
+ model fails or when the fit is not performed, e.g., for suspected
196
+ eclipsing binaries based on transit depth.
197
+
198
+ After the completion of the transit model-fitting algorithms,
199
+ the signature of the known TCE, as determined from the fitted
200
+ parameters of the all-transit fit, is removed from the light curve,
201
+ and the residual
202
+ light curve is subjected to a search for
203
+ additional planets by calling TPS in the DV component. If an
204
+ additional TCE is generated,
205
+ light curve goes
206
+ through the transit model-fitting algorithms discussed above
207
+ once again. The iterative process of the transit model-fitting and
208
+ multiple-planet search concludes when no additional TCEs are
209
+ produced or an upper limit is reached.
210
+
211
+ the residual
212
+
213
+ 3. Light-curve Preprocessing
214
+
215
+ The light curves of target stars are processed in the PDC
216
+ component before they are input in the DV component. As
217
+ described in Stumpe et al. (2012) and Smith et al. (2012),
218
+
219
+ 2
220
+
221
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
222
+
223
+ Li et al.
224
+
225
+ Figure 1. Flowchart of iterative process of transit model-fitting and multiple-planet search. In the flowchart, a rectangle represents an operation of data processing; a
226
+ diamond represents a conditional operation that determines which one of the two paths the process will take; an arrow line shows the order of operations; and an oval
227
+ represents the beginning or ending of the process.
228
+
229
+ systematic errors due to the thermal
230
+ transients and optical
231
+ distortions are estimated and compensated, outliers due to
232
+ cosmic rays and transients due to the Argabrightening events7
233
+ are removed, and sudden pixel sensitivity dropouts are
234
+ identified and corrected. Nevertheless, the PDC light curves
235
+ must be preprocessed further in preparation for transit model
236
+ fitting. The preprocessing procedures in DV include baseline
237
+ removal,
238
+ light-curve normalization, quarterly data segment
239
+ stitching, harmonic removal, and timestamp conversion.
240
+
241
+ 3.1. Baseline Removal and Light-curve Normalization
242
+
243
+ The light curve generated in the PDC component measures
244
+ −)
245
+
246
+ the brightness of the target star in units of photoelectrons (e
247
+
248
+ 7 An Argabrightening event, which was described by Witteborn et al. (2011),
249
+ is an occasional diffuse illumination of portions of the focal plane lasting a few
250
+ minutes.
251
+
252
+ per cadence.8 As the brightness of one target star is generally
253
+ measured by four different charge coupled device (CCD)
254
+ channels over the course of a year due to the quarterly rotations
255
+ of the spacecraft about the telescope boresight, the baseline of
256
+ the measured light curve of the target star varies from quarter to
257
+ quarter. The preprocessing procedure of baseline removal and
258
+ light-curve normalization removes the baseline of the measure-
259
+ ment and generates the normalized light curves so that they can
260
+ later be uniformly processed by the transit model-fitting
261
+ algorithms. This preprocessing procedure is implemented
262
+ quarter by quarter in two steps. (1) For each target star, the
263
+ median flux level is determined for each quarter. For stars on
264
+ the same CCD channel, the median flux level varies from one
265
+
266
+ 8 The flux units in the Kepler light curve files exported to the Mikulski
267
+ Archive for Space Telescopes (MAST) are actually e−/second. The conversion
268
+ from e−/cadence to e−/second is performed in the Archive (AR) component of
269
+ the pipeline.
270
+
271
+ 3
272
+
273
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
274
+
275
+ Li et al.
276
+
277
+ target star to another depending on the magnitude and spectrum
278
+ of the target star; for a given target star, the median varies from
279
+ quarter to quarter depending on the sensitivity of the CCD
280
+ pixels and the sub-pixel location of the stellar image. (2) The
281
+ median is subtracted from the corresponding quarterly light
282
+ curves and the difference is normalized by the median and
283
+ multiplied by 106, yielding a normalized light curve in units of
284
+ parts per million (ppm). For the out-of-transit data points, the
285
+ baseline value is zero. For in-transit data points, the normalized
286
+ flux is negative and its absolute value corresponds to the ratio
287
+ of the flux blocked by the transiting planet to the total flux of
288
+ the target star. For example, to an extraterrestrial observer of a
289
+ central transit, the depth of the normalized light curve of Earth
290
+ transiting the Sun is about 84×10
291
+
292
+ −6, or 84 ppm.
293
+
294
+ 3.2. Quarterly Data Segment Stitching
295
+
296
+ spacecraft
297
+
298
+ The light curve of a target star is comprised of data segments
299
+ separated by gaps that may have resulted from quarterly rolls,
300
+ monthly data downlinks, or
301
+ anomalies. The
302
+ preprocessing procedure of data segment stitching removes
303
+ the trend and transients of the light curve of the segment edges
304
+ and fills the gaps between the segments. The trend of the light
305
+ curve of each segment is identified, and the light curve at the
306
+ edges of the segments, where transients are usually observed, is
307
+ fitted with a model of linear and exponential components. Then
308
+ the detrending algorithm removes the identified trend and the
309
+ fitted components. The gaps between the data segments are
310
+ filled with different methods, depending on the length of the
311
+ gaps; the short gaps are filled with an auto-regressive model
312
+ and the long gaps are filled via data reflection and tapering
313
+ (Jenkins 2002; Jenkins et al. 2010b, 2017).
314
+
315
+ 3.3. Harmonic Removal
316
+ The harmonic removal procedure identifies and removes
317
+ sinusoidal harmonic components, which are significant in the
318
+ light curves of target stars such as rotating and contact binaries.
319
+ The light curve is first processed with a Fast Fourier Transform
320
+ (FFT) to determine the frequencies of the significant harmonic
321
+ components. Then the magnitude and phase of the components
322
+ are fitted and the significant harmonic components are removed
323
+ from the light curve (Jenkins 2002; Jenkins et al. 2010b, 2017).
324
+ It is possible that the harmonic removal process may degrade
325
+ the transits of short-period planets. This was discussed by
326
+ Christiansen et al. (2013, 2015).
327
+
328
+ 3.4. Timestamp Conversion
329
+ Based on Kepler’s laws of planetary motion, the transits of
330
+ exoplanets are inherently periodic9 if the observer is located at
331
+ the barycenter of the solar system and the events are measured
332
+
333
+ 9 This neglects transit-timing variations, which can be quite large for
334
+ dynamically packed planetary systems with planets in near-orbital resonances.
335
+
336
+ Table 1
337
+ Time Frames and Timestamps Before and After the Timestamp Conversion
338
+
339
+ Before/After Conversion
340
+
341
+ Time Frame
342
+
343
+ Timestamp
344
+
345
+ Before
346
+ After
347
+
348
+ UTC
349
+ TDB
350
+
351
+ MJD
352
+ BKJD (=BMJD-54,832.5)
353
+
354
+ in the barycentric dynamical time (TDB) frame. However, the
355
+ timestamps associated with the PDC light curves provided to
356
+ DV are modified Julian day (MJD) numbers, which correspond
357
+ to the time when the light of the target star arrives at the Kepler
358
+ spacecraft in the Coordinated universal Time (UTC) frame.
359
+ Before the transit model-fitting algorithms are applied,
360
+ barycentric time corrections are applied to obtain timestamps
361
+ in barycentric modified Julian day (BMJD) numbers,
362
+ to
363
+ correspond to the time when the light from the events of the
364
+ target star would arrive at the barycenter of the solar system in
365
+ the TDB frame.
366
+
367
+ The algorithm to determine each BMJD timestamp requires
368
+ the following inputs. The ephemeris of the Kepler spacecraft,
369
+ the ephemeris of the solar system, and the celestial coordinates
370
+ of the target star. Then the difference between the time when
371
+ the light of the events of the target star would have reached the
372
+ barycenter of the solar system and the time when the same light
373
+ arrived at the Kepler spacecraft, which is located in an Earth-
374
+ trailing heliocentric orbit, is calculated. Finally, the BMJD
375
+ timestamps are determined as the sum of the MJD timestamps
376
+ and the aforementioned barycentric time corrections. To
377
+ simplify the processing and storage of the Kepler science data,
378
+ a new timestamp, barycentric Kepler Julian date (BKJD), is
379
+ defined and used in the Kepler Science Data Processing
380
+ Pipeline and the NASA Exoplanet Archive. By definition,
381
+ BKJD is equal to BMJD minus a constant of 54,832.5 days,
382
+ which corresponds to 12:00:00 noon on 2009 January 1 (the
383
+ first day of the year when the Kepler spacecraft was launched).
384
+ After the preprocessing procedure of timestamp conversion, all
385
+ light curves are associated with BKJD timestamps. The time
386
+ timestamp
387
+ frames and the timestamps before and after
388
+ conversion in the preprocessing are summarized in Table 1.
389
+
390
+ As an example, Figure 2 shows two segments of the light
391
+ curve of the target star KIC 8478994, or Kepler-37, before and
392
+ after the preprocessing procedures. As illustrated in the figure,
393
+ the light curve before the preprocessing shows the absolute flux
394
+ value in units of photoelectrons, timestamped in MJD, and the
395
+ light curve after the preprocessing shows the dimensionless
396
+ normalized flux value, timestamped in BKJD.
397
+
398
+ 4. Geometric Transit Model
399
+ The transit model-fitting algorithms of the DV component
400
+ employ the geometric transit model of Mandel & Agol (2002),
401
+ including a nonlinear limb-darkening model, parameterized as
402
+
403
+ 4
404
+
405
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
406
+
407
+ Li et al.
408
+
409
+ Figure 2. Flux timeseries of KIC 8478994 before (top) and after (bottom) the preprocessing procedures.
410
+
411
+ per Claret & Bloemen (2011). The limb-darkening depends on
412
+ the stellar parameters, such as radius, Rs (solar radii), surface
413
+ log (log10(cm s−2)), metallicity, log10[M/H] (dimen-
414
+ gravity,
415
+ sionless), and effective temperature, Teff
416
+ (K), which are
417
+ extracted from the Kepler Input Catalog (KIC) (Brown et al.
418
+ 2011) or override to KIC parameter values (Mathur et al. 2017).
419
+
420
+ g
421
+
422
+ 4.1. Fitted Parameters
423
+
424
+ In the geometric transit model,
425
+
426
+ the eccentricity and the
427
+ longitude of periapsis of the planet orbit around the host star
428
+ are assumed to be 0, and the five parameters to be fitted are
429
+ listed below.
430
+
431
+ (cid:129) Transit epoch time, tepoch (BKJD): the time corresponding
432
+
433
+ to the center of the first detected transit;
434
+
435
+ (cid:129) Orbital period, P (days): the interval between consecutive
436
+ planetary transits, i.e., the period of the planet’s orbit;
437
+ (cid:129) Ratio of planet radius to stellar radius, Rp/Rs (dimension-
438
+ less): the ratio of the planet radius divided by the stellar
439
+ radius;
440
+
441
+ (cid:129) Ratio of semimajor axis to stellar radius, a/Rs (dimension-
442
+ less): as the eccentricity of the planet’s orbit is assumed to
443
+
444
+ be zero, this is the ratio of the distance between the planet
445
+ and the host star divided by the stellar radius;
446
+
447
+ (cid:129) Impact parameter, b (dimensionless):
448
+
449
+ the sky-projected
450
+ distance between the center of the stellar disk and the
451
+ center of the planet disk at conjunction, normalized by the
452
+ stellar radius.
453
+
454
+ As discussed in Section 6.1.2,
455
+
456
+ the fitted parameters are
457
+ (LM)
458
+ determined with the iterative Levenberg-Marquardt
459
+ algorithm (Levenberg 1944; Marquardt 1963). The all-transit
460
+ fit and odd–even transit fit are both initialized with the fitted
461
+ parameters of the reduced-parameter fit with the minimum χ2
462
+ metric.
463
+
464
+ In the reduced-parameter fits, the impact parameter b is set to
465
+ fixed values of 0.1, 0.3, 0.5, 0.7, and 0.9. The initial values of
466
+ the fitted parameters tepoch, P, Rp/Rs, and a/Rs are determined
467
+ from the TCE parameters provided by the TPS component. The
468
+ TPS value for orbital period can be used directly. Note that the
469
+ transit epoch time from the TPS component is in units of MJD,
470
+ while the fitted parameter of tepoch is in units of BKJD;
471
+ therefore, a unit conversion is required. The initial values of
472
+ Rp/Rs and a/Rs are determined from the TCE parameters
473
+
474
+ 5
475
+
476
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
477
+
478
+ Li et al.
479
+
480
+ according to
481
+
482
+ ⎛
483
+ ⎜
484
+ ⎝
485
+
486
+ =
487
+
488
+ R
489
+
490
+ p
491
+ R
492
+ s
493
+
494
+ TPS
495
+
496
+ SES
497
+ r
498
+
499
+ flux
500
+
501
+ 1
502
+ 2
503
+
504
+ ⎞
505
+ ⎟
506
+ ⎠
507
+
508
+ ⎛
509
+ ⎜
510
+ ⎝
511
+
512
+ d
513
+
514
+ lc
515
+
516
+ d
517
+
518
+ tr TPS
519
+
520
+ ,
521
+
522
+ 1
523
+ 4
524
+
525
+ ⎞
526
+ ⎟
527
+ ⎠
528
+
529
+ and
530
+
531
+ a
532
+ R
533
+ s
534
+
535
+ =
536
+
537
+ ⎛
538
+ ⎜
539
+ ⎜
540
+ ⎜
541
+ ⎝
542
+
543
+ (
544
+
545
+ 1
546
+
547
+ +
548
+
549
+ R
550
+
551
+ p
552
+ R
553
+ s
554
+
555
+ 2
556
+
557
+ )
558
+
559
+ -
560
+
561
+ 2
562
+
563
+ b
564
+
565
+ 2
566
+ sin
567
+
568
+ (
569
+
570
+ p
571
+
572
+ d
573
+ 24
574
+
575
+ tr TPS
576
+ ,
577
+ P
578
+
579
+ TPS
580
+
581
+ )
582
+
583
+ +
584
+
585
+ 2
586
+
587
+ b
588
+
589
+ 1
590
+ 2
591
+
592
+ ⎞
593
+ ⎟
594
+ ⎟
595
+ ⎟
596
+ ⎠
597
+
598
+ ,
599
+
600
+ ( )
601
+ 1
602
+
603
+ ( )
604
+ 2
605
+
606
+ where the single event statistic, SESTPS (dimensionless), orbital
607
+ period, PTPS (days), and transit duration, dtr, TPS (hours) are
608
+ TCE parameters determined in the TPS component. rflux is the
609
+ ratio of the light-curve value to the uncertainty. dlc (hours) is
610
+ the duration of a long-cadence (LC) interval (29.4 min or
611
+ 0.49 hr).
612
+
613
+ 4.2. Derived Parameters
614
+ Once the transit model-fitting algorithm has converged,
615
+ the
616
+
617
+ several additional parameters regarding the planet or
618
+ transits can be derived from the fitted parameters.
619
+
620
+ Figure 3. Schematic of planetary transit and associated light curve with the
621
+ depth, duration, ingress time, and epoch time indicated.
622
+
623
+ Given the stellar
624
+
625
+ radius, Rs,
626
+ determined directly from the fitted parameter, Rp/Rs:
627
+
628
+ the planet
629
+
630
+ radius, Rp,
631
+
632
+ R
633
+
634
+ p
635
+
636
+ =
637
+
638
+ ⎛
639
+ ⎜
640
+ ⎝
641
+
642
+ ⎞
643
+ ⎟
644
+ ⎠
645
+
646
+ ⎛
647
+ ⎜
648
+ ⎝
649
+
650
+ ⎞
651
+ ⎟
652
+ ⎠
653
+
654
+ R
655
+
656
+ p
657
+ R
658
+ s
659
+
660
+ R
661
+ R
662
+
663
+
664
+
665
+ Å
666
+
667
+ R ,
668
+ s
669
+
670
+ is
671
+
672
+ ( )
673
+ 3
674
+
675
+ The inclination of the planet orbit, i, in units of degrees, is
676
+
677
+ determined from fitted parameters b and a/Rs:
678
+
679
+ i
680
+
681
+ =
682
+
683
+ 180
684
+ p
685
+
686
+ -
687
+
688
+ cos
689
+
690
+ 1
691
+
692
+ ⎛
693
+ ⎜
694
+ ⎝
695
+
696
+ b
697
+ a R
698
+ s
699
+
700
+ ⎞
701
+ ⎟
702
+ ⎠
703
+
704
+ .
705
+
706
+ ( )
707
+ 6
708
+
709
+ where Re and R⊕ are radii of the Sun and Earth, respectively,
710
+ both in units of m. Because Rs is in units of solar radii, Rp,
711
+ given by Equation (3), is in units of Earth radii.
712
+
713
+ Before calculating the semimajor axis of the planet orbit, a,
714
+ the planet–star separation for a circular orbit, the acceleration
715
+ due to gravity at the surface of the star, g, should first be
716
+ determined from the stellar parameter
717
+
718
+ log
719
+
720
+ as
721
+
722
+ g
723
+
724
+ g
725
+
726
+ =
727
+
728
+ 1
729
+ 100
730
+
731
+ 10
732
+
733
+ log
734
+
735
+ g
736
+
737
+ .
738
+
739
+ ( )
740
+ 4
741
+
742
+ log
743
+
744
+ A factor of 1/100 is required to convert acceleration g to units
745
+ of m s−2 from
746
+
747
+ in units of log10(cm s−2).
748
+ The semimajor axis of the planetary orbit a is not determined
749
+ directly from the fitted parameter a/Rs, but derived from the
750
+ fitted orbital period, P, based on Kepler’s third law:
751
+
752
+ g
753
+
754
+ the transit depth, D,
755
+
756
+ As illustrated in Figure 3,
757
+
758
+ transit
759
+ duration, dtr, and transit ingress time, din, are another set of
760
+ parameters defining the size and shape of a transit. The transit
761
+ depth, D, is determined as the absolute value of the minimum
762
+ the normalized light curve generated from the fitted
763
+ of
764
+ parameters (to be discussed in Section 5), multiplied by a
765
+ factor of 106 to convert the dimensionless normalized flux
766
+ value to the transit depth in units of ppm. The parameters dtr
767
+ and din, both in units of hours, are derived from fitted
768
+ parameters Rp/Rs, a/Rs, b, and P with the following equations:
769
+ (
770
+
771
+ )
772
+
773
+ -
774
+
775
+ +
776
+
777
+ P
778
+
779
+ 1
780
+
781
+ b
782
+
783
+ R
784
+
785
+ 2
786
+
787
+ 2
788
+
789
+ , and
790
+
791
+ ( )
792
+ 7
793
+
794
+ d
795
+
796
+ tr
797
+
798
+ =
799
+
800
+ 24
801
+ p
802
+
803
+ -
804
+
805
+ 1
806
+
807
+ sin
808
+
809
+ p
810
+ R
811
+ s
812
+ 2
813
+
814
+ ( )
815
+
816
+ a
817
+ R
818
+ s
819
+
820
+ -
821
+
822
+ 2
823
+
824
+ b
825
+
826
+ ⎛
827
+ ⎜
828
+ ⎜
829
+ ⎜⎜
830
+ ⎝
831
+
832
+ ⎞
833
+ ⎟
834
+ ⎟
835
+ ⎟⎟
836
+ ⎠
837
+
838
+ d
839
+
840
+ in
841
+
842
+ (
843
+
844
+ 86400
845
+
846
+ a
847
+
848
+ =
849
+
850
+ ⎡
851
+ ⎢
852
+ ⎣
853
+
854
+ 1
855
+ f
856
+ au
857
+
858
+ s
859
+
860
+ P R R
861
+ )(
862
+ p
863
+
864
+ 2
865
+
866
+ )
867
+
868
+
869
+
870
+ g
871
+
872
+ 2
873
+ 3
874
+
875
+ ⎤
876
+ ⎥
877
+ ⎦
878
+
879
+ ,
880
+
881
+ ( )
882
+ 5
883
+
884
+ =
885
+
886
+ P
887
+
888
+ 12
889
+ p
890
+
891
+ -
892
+
893
+ sin
894
+
895
+ 1
896
+
897
+ where fau is the factor to convert the astronomical unit (au) to m
898
+ (i.e., the number of m in one au). Please note P and Rs are in
899
+ units of days and solar radii, respectively, so Equation (5) gives
900
+ the semimajor axis of the planet orbit, a, in au.
901
+
902
+ ⎛
903
+ ⎜
904
+ ⎜
905
+ ⎜⎜
906
+ ⎝
907
+
908
+ (
909
+
910
+ 1
911
+
912
+ +
913
+
914
+ ( )
915
+
916
+ a
917
+ R
918
+ s
919
+
920
+ R
921
+
922
+ p
923
+ R
924
+ s
925
+ 2
926
+
927
+ 2
928
+
929
+ )
930
+
931
+ -
932
+
933
+ 2
934
+
935
+ b
936
+
937
+ -
938
+
939
+ 2
940
+
941
+ b
942
+
943
+ -
944
+
945
+ (
946
+
947
+ 1
948
+
949
+ -
950
+
951
+ ( )
952
+
953
+ a
954
+ R
955
+ s
956
+
957
+ R
958
+
959
+ p
960
+ R
961
+ s
962
+ 2
963
+
964
+ 2
965
+
966
+ )
967
+
968
+ -
969
+
970
+ 2
971
+
972
+ b
973
+
974
+ -
975
+
976
+ 2
977
+
978
+ b
979
+
980
+ ⎞
981
+ ⎟
982
+ ⎟
983
+ ⎟⎟
984
+ ⎠
985
+
986
+ .
987
+
988
+ ( )
989
+ 8
990
+
991
+ The planet equilibrium temperature, Teq, an estimate of the
992
+ surface temperature of the planet, is calculated assuming a
993
+
994
+ 6
995
+
996
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
997
+
998
+ Li et al.
999
+
1000
+ Table 2
1001
+ Fitted and Derived Parameters of the Transit Model
1002
+
1003
+ Parameter
1004
+
1005
+ Symbol
1006
+
1007
+ Unit
1008
+
1009
+ Fitted/Derived
1010
+
1011
+ Transit epoch time
1012
+ Planet orbital period
1013
+ ratio of planet radius to star
1014
+
1015
+ radius
1016
+
1017
+ Ratio of semimajor axis to
1018
+
1019
+ star radius
1020
+ Impact parameter
1021
+ Planet radius
1022
+ Planet orbit semimajor axis
1023
+ Planet orbit inclination
1024
+ Transit depth
1025
+ Transit duration
1026
+ Transit ingress time
1027
+ Planet equilibrium
1028
+ temperature
1029
+
1030
+ Planet effective stellar flux
1031
+
1032
+ tepoch
1033
+ P
1034
+ Rp/Rs
1035
+
1036
+ a/Rs
1037
+
1038
+ b
1039
+ Rp
1040
+ a
1041
+ i
1042
+ D
1043
+ dtr
1044
+ din
1045
+ Teq
1046
+
1047
+ feff
1048
+
1049
+ BKJD
1050
+ Day
1051
+ Dimensionless
1052
+
1053
+ Fitted
1054
+ Fitted
1055
+ Fitted
1056
+
1057
+ Dimensionless
1058
+
1059
+ Fitted
1060
+
1061
+ Dimensionless
1062
+ Earth radius
1063
+ au
1064
+ Degree
1065
+ ppm
1066
+ Hour
1067
+ Hour
1068
+ K
1069
+
1070
+ Fitted
1071
+ Derived
1072
+ Derived
1073
+ Derived
1074
+ Derived
1075
+ Derived
1076
+ Derived
1077
+ Derived
1078
+
1079
+ Dimensionless
1080
+
1081
+ Derived
1082
+
1083
+ thermodynamic equilibrium is reached between the incident
1084
+ stellar flux and the radiated heat from the planet:
1085
+
1086
+ T
1087
+ eq
1088
+
1089
+ =
1090
+
1091
+ T
1092
+ eff
1093
+
1094
+ (
1095
+
1096
+ 1
1097
+
1098
+ -
1099
+
1100
+ 1
1101
+ 4a
1102
+ )
1103
+
1104
+ R R
1105
+ s
1106
+ a f
1107
+ 2
1108
+
1109
+ au
1110
+
1111
+
1112
+
1113
+ ,
1114
+
1115
+ ( )
1116
+ 9
1117
+
1118
+ where a is the semimajor axis of the planetary orbit in au
1119
+ determined by Equation (5), α is the albedo of the planet,
1120
+ whose default value is set to 0.3, and both Teff and Teq are in K.
1121
+ The planet effective stellar flux, feff, defined as the ratio of
1122
+ the flux of the host star at the top of planet’s atmosphere to the
1123
+ solar flux at the top of Earth’s atmosphere, is calculated as
1124
+ ⎛
1125
+ ⎜
1126
+ ⎝
1127
+
1128
+ f =
1129
+ eff
1130
+
1131
+ ⎞
1132
+ 2
1133
+ ⎟
1134
+ ⎠
1135
+
1136
+ ⎞
1137
+ ⎟
1138
+ ⎠
1139
+
1140
+ T
1141
+ eff
1142
+
1143
+ ⎛
1144
+ ⎜
1145
+ ⎝
1146
+
1147
+ 10
1148
+
1149
+ (
1150
+
1151
+ ,
1152
+
1153
+ 4
1154
+
1155
+ )
1156
+
1157
+ R
1158
+ s
1159
+ a
1160
+
1161
+ T
1162
+ eff
1163
+
1164
+ ,
1165
+
1166
+
1167
+
1168
+ where a is determined by Equation (5), and Teff,e is the
1169
+ effective temperature of the Sun in units of K.
1170
+
1171
+ The fitted and derived parameters of the transit model are
1172
+
1173
+ summarized in Table 2.
1174
+
1175
+ 5. Geometric Transit Signal Generator
1176
+
1177
+ The geometric transit signal generator generates a light curve
1178
+ at an array of cadence timestamps in BKJD (nominally the
1179
+ timestamps corresponding to the midpoints of cadences) with
1180
+ the fitted parameters of a geometric transit model described in
1181
+ Section 4.1. The coefficients of the limb-darkening model are
1182
+ determined by the stellar parameters of the target star (Claret &
1183
+ Bloemen 2011).
1184
+
1185
+ The computation of the light curve is implemented in the
1186
+ following steps. First, an array of oversampled timestamps is
1187
+ timestamps. This is
1188
+ constructed from the input array of
1189
+ necessary to obtain an accurate flux level estimate at
1190
+ the
1191
+ temporal resolution of the data (29.4 min). For each element in
1192
+
1193
+ 7
1194
+
1195
+ the input array of timestamps, a subarray of 11 oversampled
1196
+ timestamps is generated. The step size of the oversampled
1197
+ timestamps is 1/11 of a LC interval, or 2.67 min. The center
1198
+ element of the subarray, the sixth of the 11 elements, is equal to
1199
+ the corresponding element in the input array of timestamps.
1200
+ The oversampled timestamps that fall within a given transit
1201
+ (including a small buffer on each side of the transit) are
1202
+ identified with the parameters tepoch and P. A circular Keplerian
1203
+ orbit, normalized by the stellar radius, Rs, is determined from
1204
+ the parameters a/Rs and b. The position vectors of the planet in
1205
+ the orbit are computed and the corresponding impact
1206
+ parameters are determined by projecting the position vectors
1207
+ to the plane perpendicular to the direction of the target star. The
1208
+ relative flux value, the ratio of the stellar flux blocked by the
1209
+ transiting planet to the unblocked stellar flux, is calculated for
1210
+ each oversampled timestamp with the impact parameter b, the
1211
+ fitted parameter Rp/Rs, and the limb-darkening coefficients.
1212
+ Finally, the relative flux at each of the input timestamps is
1213
+ determined as the mean of 11 relative flux values at
1214
+ the
1215
+ corresponding oversampled timestamps.
1216
+
1217
+ Figure 4 shows the normalized flux timeseries generated
1218
+ by the geometric transit signal generator with following
1219
+ tepoch=138.50000 days, P=10.30405 days,
1220
+ parameters:
1221
+ Rp/Rs=0.0155697, a/Rs=18.7471, and b=0.1, which
1222
+ are determined by the reduced-parameter fit (to be discussed in
1223
+ Section 6.2) of the sixth TCE of the target star KIC 6541920,
1224
+ also known as the planet Kepler-11b. As shown in the figure,
1225
+ the step size of the input timestamps is the duration of a LC
1226
+ (29.4 min), and the normalized flux values at
1227
+ the input
1228
+ and oversampled timestamps are plotted in red and blue,
1229
+ respectively.
1230
+
1231
+ Because the surface brightness of a star appears to vary due
1232
+ to the limb-darkening effect, the calculation of the normalized
1233
+ flux value is
1234
+ implemented with an iterative numerical
1235
+ integration algorithm. At each iteration, the integration step is
1236
+ in half and an updated normalized flux timeseries is
1237
+ cut
1238
+ determined with the nonlinear limb-darkening model. The
1239
+ iterative process is terminated when the required precision is
1240
+ satisfied or when an upper limit of the execution time of the
1241
+ iterative algorithm is reached. If the parameter Rp/Rs is less
1242
+ than 0.01, a small-body approximation is used to speed up the
1243
+ algorithm, assuming the stellar surface brightness is constant
1244
+ under the disk of the eclipsing object (Mandel & Agol 2002).
1245
+ The five fitted parameters defined in Section 4.1 can be
1246
+ divided in two relatively independent groups: (1) the transit
1247
+ tepoch, and the orbital period, P, define the
1248
+ epoch time,
1249
+ occurrence time of the transits; (2) the ratio of planet radius
1250
+ to star radius, Rp/Rs, the ratio of semimajor axis to star radius,
1251
+ a/Rs, and the impact parameter, b, define the depth, duration,
1252
+ and shape of the transits.
1253
+
1254
+ Figure 5 illustrates how the variations of the parameters
1255
+ Rp/Rs, a/Rs, and b change the depth, duration, and shape of the
1256
+ transits in the light curves. As the reference for comparison, the
1257
+
1258
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1259
+
1260
+ Li et al.
1261
+
1262
+ Figure 4. Normalized light curve of the sixth TCE of KIC 6541920 generated by the geometric transit signal generator. The normalized flux values at the oversampled
1263
+ timestamps, whose step size is 1/11 of a LC interval, or 2.67 min, are plotted in blue. Each of the normalized flux values at the LC interval, or 29.4 min, is determined
1264
+ as the mean of 11 corresponding values at oversampled timestamps and plotted in red.
1265
+
1266
+ the corresponding model
1267
+
1268
+ light curve shown in Figure 4 is plotted as blue in Figure 5. The
1269
+ corresponding parameter values are used as references for the
1270
+ parameter variations. When Rp/Rs is increased by 10% and
1271
+ 20% to its reference value,
1272
+ light
1273
+ curves are plotted as red and black lines, respectively, in the
1274
+ plot on the top of Figure 5. Because Rp/Rs defines the relative
1275
+ size of the transiting planet to the host star, an increase of
1276
+ Rp/Rs, meaning more stellar flux is blocked by the transiting
1277
+ planet, leading to an increase of the transit depth. When a/Rs is
1278
+ the
1279
+ increased by 10% and 20% to its reference value,
1280
+ corresponding model light curves are plotted as red and black
1281
+ lines, respectively, in the middle plot of Figure 5. Because the
1282
+ orbital period, P, is fixed, an increase of a/Rs, indicating a
1283
+ decrease of the stellar radius, Rs, leads to a decrease of the
1284
+ transit duration. When b is changed to 0.3, 0.5, 0.7, and 0.9, the
1285
+ corresponding model light curves are plotted as red, black,
1286
+ magenta, and green lines, respectively,
1287
+ in the plot on the
1288
+ bottom of Figure 5. An increase of b moves the transit
1289
+ trajectory toward the edge of the stellar disk and results in a
1290
+ decrease of the transit duration. Because the point at the center
1291
+ of the transit moves toward the edge of the stellar disk, the
1292
+ transit depth decreases as well due to the limb-darkening effect.
1293
+
1294
+ 6. Geometric Model-fitting Algorithms
1295
+ The inputs of the geometric model-fitting algorithms include
1296
+ (1) the light curve after the preprocessing procedures described
1297
+ in Section 3, and (2) the TCE parameters, including transit
1298
+ epoch time, orbital period, trial transit duration, and single and
1299
+ multiple event statistics, generated by the TPS component
1300
+ (Jenkins 2002; Jenkins et al. 2010b, 2017; Tenenbaum et al.
1301
+ 2012, 2013, 2014; Seader et al. 2015; Twicken et al. 2016).
1302
+
1303
+ Section 6.1 discusses an iterative whitening and model-
1304
+ fitting process, used in the all-transit fit, odd–even transit fit,
1305
+ and reduced-parameter fit. The TCE parameters are used to
1306
+ seed the initial values of reduced-parameter fits. Section 6.2
1307
+ describes a fitting algorithm to resolve the degenerate problem
1308
+ of fitting the impact parameters. Section 6.3 describes the
1309
+ algorithms to fit odd and even transits, whose outputs are used
1310
+ in the diagnostic test to distinguish transiting planets from
1311
+ circular eclipsing binaries that have been detected at one-half of
1312
+ their true orbital period (Twicken et al. 2018).
1313
+
1314
+ The fitter outputs, which are generated when the fitting
1315
+ algorithm completes successfully, are described in Section 6.4.
1316
+ The alert messages, which are generated when the fitting
1317
+ algorithm fails, are discussed in Section 6.5.
1318
+
1319
+ 6.1. Iterative Whitening and Model Fitting
1320
+
1321
+ Compared with transit features, secular variations due to
1322
+ pointing drift, focus variations, and stellar variability can be
1323
+ quite large. Secular variations of the light curve, appearing as
1324
+ correlated noise, can lead to biases in the fitted parameters of
1325
+ the geometric transit model. Therefore, a whitening filter is
1326
+ applied to the light curves before transit model fitting to
1327
+ account explicitly for the correlation structure of the noise.
1328
+ Considering that the whitening filter changes the shape of the
1329
+ transits, the same whitening filter is applied to the model light
1330
+ curve generated by the geometric transit signal generator.
1331
+
1332
+ The flux timeseries of a target star at times ti, i=1, 2, KN,
1333
+ is denoted as y(ti). Let q denote a 5×1 vector of fitted
1334
+ parameters:
1335
+
1336
+ ⎤⎦
1337
+ T
1338
+ P R R a R b .
1339
+
1340
+ p
1341
+
1342
+ s
1343
+
1344
+ s
1345
+
1346
+ (
1347
+
1348
+ 11
1349
+
1350
+ )
1351
+
1352
+ q = ⎡⎣
1353
+
1354
+ t
1355
+ epoch
1356
+
1357
+ 8
1358
+
1359
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1360
+
1361
+ Li et al.
1362
+
1363
+ Figure 5. Light curves generated by the geometric transit signal generator with different parameters of Rp/Rs (top), a/Rs (middle), and b (bottom). See the text for an
1364
+ explanation.
1365
+
1366
+ The predicted light curve generated from the geometric transit
1367
+ model with the parameter vector q is denoted as s t ,i q(
1368
+ ). When
1369
+ the whitening filter is applied to the timeseries y(ti) and s t ,i q(
1370
+ ),
1371
+
1372
+ the corresponding whitened timeseries are denoted as y ti˜( ) and
1373
+ ), respectively. The geometric transit model fitting is
1374
+ s t ,i q
1375
+ ˜(
1376
+ implemented with a LM algorithm to search for the vector q in
1377
+
1378
+ 9
1379
+
1380
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1381
+
1382
+ Li et al.
1383
+
1384
+ Figure 6. Block diagram of the iterative whitening and model-fitting process. Two loops are shown in the figure: the outer loop, shown in the rectangle of dashed
1385
+ lines, includes a whitening filter and a transit model fitter, and the parameters of the whitening filter are updated on each iteration with the residuals of the transit model
1386
+ fitter; the inner loop, shown in the area surrounded by dotted lines, includes the LM fit and robust weight reassignment.
1387
+
1388
+ the parameter space to minimize the following weighted
1389
+ nonlinear least-squares cost function:
1390
+
1391
+ 2
1392
+
1393
+ c
1394
+
1395
+ N
1396
+ åq
1397
+ =
1398
+ ( )
1399
+ =
1400
+ 1
1401
+
1402
+ i
1403
+
1404
+ w y t
1405
+ [ ˜( )
1406
+ i
1407
+
1408
+ i
1409
+
1410
+ -
1411
+
1412
+ s t ,
1413
+ ˜ (
1414
+ i
1415
+
1416
+ q
1417
+
1418
+ 2
1419
+ )]
1420
+
1421
+ ,
1422
+
1423
+ (
1424
+
1425
+ 12
1426
+
1427
+ )
1428
+
1429
+ where wi, i=1, 2, KN are weights, ranging between 0 and 1.
1430
+ During the fit, these robust weights are adjusted to deemphasize
1431
+ points with large departures from the model values, to reduce
1432
+ the impact of outliers (Holland & Welsch 1977).
1433
+
1434
+ Let y˜ and s q˜( ) denote vectors of measured and predicted
1435
+ light curves in the whitened domain, respectively, and W
1436
+ denote a diagonal matrix of the weights. Equation (12) can be
1437
+ rewritten in the following matrix form:
1438
+
1439
+ c
1440
+
1441
+ 2 q
1442
+ ( )
1443
+
1444
+ = -
1445
+ y
1446
+ [˜
1447
+
1448
+ T
1449
+ ˜( )]
1450
+ q
1451
+ s
1452
+
1453
+ W y
1454
+ [˜
1455
+
1456
+ -
1457
+
1458
+ ˜( )]
1459
+ q
1460
+ s
1461
+
1462
+ .
1463
+
1464
+ (
1465
+
1466
+ 13
1467
+
1468
+ )
1469
+
1470
+ Because the out-of-transit
1471
+
1472
+ light-curve data just show the
1473
+ measurement noise around the baseline value of zero, they
1474
+ offer no information to characterize the transits. Therefore, the
1475
+ transit model fitting is restricted to the data within or close
1476
+ to the transits. The center times of the transits are predicted
1477
+ from the parameters tepoch and P, and only the light curve data
1478
+ whose timestamps fall in the time ranges of five times the
1479
+ transit duration, centered at the transit center, are used in the
1480
+ geometric transit model fit. The data selection can also be
1481
+ viewed as a model fit implemented with different weights to all
1482
+
1483
+ data points; the weight is set to 1 in Equation (12) when the
1484
+ difference between the timestamp of the data point and the
1485
+ center time of the nearest transit is less than 2.5 times the transit
1486
+ duration. Otherwise, the weight is set to 0.
1487
+
1488
+ For each TCE generated by TPS, the geometric transit model
1489
+ fitting is implemented with a loop that includes an adaptive
1490
+ whitening filter and a robust LM transit fitter, as shown in
1491
+ Figure 6. The whitening filter transforms the correlated noise in
1492
+ the measured flux timeseries to uncorrelated, or white, noise.
1493
+ The predicted light curve is subjected to the same whitening
1494
+ filter, so the fitted parameters of the geometric transit model are
1495
+ determined by nonlinear least-squares fitting in the whitened
1496
+ domain. The fit residual is utilized to update the parameters of
1497
+ the whitening filter on each iteration. Robust weights are
1498
+ assigned to each point of the flux timeseries so that data with
1499
+ large errors are assigned small weights in the nonlinear least-
1500
+ squares fitting algorithm. The iterative whitening and fitting
1501
+ loop is terminated when both the whitening filter and the transit
1502
+ fitter converge or a predefined iteration limit is reached.
1503
+
1504
+ 6.1.1. Whitening Filter
1505
+
1506
+ Considering the non-stationary nature of the stellar varia-
1507
+ bility, an adaptive whitening filter is generated and used to
1508
+ remove variations in the light curve.
1509
+
1510
+ 10
1511
+
1512
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1513
+
1514
+ Li et al.
1515
+
1516
+ Figure 7. Whitened flux timeseries of KIC 8478994.
1517
+
1518
+ The whitening filter is implemented in the following steps.
1519
+ (1) The flux timeseries and the model transit light curves are
1520
+ mapped into a two dimensional array of whitening coefficients,
1521
+ localizing the signal both in time and frequency with the
1522
+ Overcomplete Wavelet Transform (OWT), a modified version
1523
+ of the discrete wavelet transform (Jenkins 2002; Jenkins et al.
1524
+ 2010b, 2017). (2) The noise power in each wavelet bandpass is
1525
+ estimated using a moving median absolute deviation (MAD)
1526
+ filter. (3) The wavelet coefficients of the flux timeseries and the
1527
+ model transit light curve are normalized by the root-mean-
1528
+ square (rms) noise power estimates. Finally,
1529
+ the whitened
1530
+ timeseries is reconstructed from the updated wavelet coeffi-
1531
+ cients with the inverse OWT.
1532
+
1533
+ Figure 7 shows the whitened flux timeseries of KIC 8478994
1534
+ in an interval of 100 days, which is produced when the
1535
+ unwhitened normalized flux timeseries, shown earlier on the
1536
+ lower panel of Figure 2, is processed with a whitening filter.
1537
+ Figure 8 illustrates the same unwhitened and whitened flux
1538
+ timeseries in an interval of six days; the distortion of the
1539
+ whitening filter on the shape of the transit is evident. It can be
1540
+ seen that the depth of the transit is approximately 6×10−4, or
1541
+ 600 ppm, while stellar variability produces variations of more
1542
+ than 3×10−4 in the unwhitened flux timeseries. The whitened
1543
+ flux timeseries, whose standard deviation is equal to 1, is in
1544
+ units of standard deviations of the unwhitened flux.
1545
+
1546
+ model-fitting algorithms, all of the fitted parameters are forced
1547
+ to be positive values. When an updated value of a parameter is
1548
+ negative in the search process, the parameter is set to the
1549
+ absolute value of the updated value so that all fitted parameters
1550
+ are positive.
1551
+
1552
+ An additional subtlety to the parameterization is that the
1553
+ impact parameter is constrained to lie in the range [0, 1], but the
1554
+ LM algorithm implicitly requires all fit parameters to be valid
1555
+ over all real values. To address this mismatch, a nonlinear
1556
+ transformation in the form of a sin function is performed
1557
+ between the “internal” parameter used by the LM algorithm
1558
+ and the “external parameter” used in the geometric transit
1559
+ ] used by
1560
+ model; this transformation maps the range
1561
+ the LM algorithm to the range [−1, 1] for the impact parameter
1562
+ in the geometric transit model. Negative values are also
1563
+ updated with the corresponding absolute values, as discussed
1564
+ above.
1565
+
1566
+ ,-¥ ¥
1567
+
1568
+ [
1569
+
1570
+ In the DR25 processing with SOC 9.3 codebase, the iterative
1571
+ search process for the parameter vector q halted if the relative
1572
+ variation of the χ2 metric was less than 0.1%, or the absolute
1573
+ value of the difference of the fitted parameters was less than
1574
+ 10% of the corresponding uncertainties, or a preset limit of 100
1575
+ iterations was reached. The threshold values are configurable
1576
+ DV parameters.
1577
+
1578
+ 6.1.2. LM Fit of Geometric Transit Model Parameters
1579
+
1580
+ The LM algorithm is employed to search for parameters that
1581
+ minimize the nonlinear least-squares cost function defined in
1582
+ Equations (12) and (13).
1583
+
1584
+ In the general form of the LM algorithm,
1585
+
1586
+ there is no
1587
+ restriction on the values of the fitted parameters. However, in
1588
+ the geometric transit model, the parameters P, Rp/Rs, a/Rs,
1589
+ and b must be positive. Therefore, in the geometric transit
1590
+
1591
+ 6.1.3. Robust Fit
1592
+ In the weighted nonlinear least-squares fitting problem given
1593
+ by Equations (12) or (13), the weight of each data point used in
1594
+ the fit is initialized to either 1 or 0, depending on whether the
1595
+ timestamp of the data point is within 2.5 times the transit
1596
+ duration from the center time of the nearest transit. However,
1597
+ when some of the selected data points are outliers, the fitting
1598
+ algorithm converges to a compromised solution between the
1599
+
1600
+ 11
1601
+
1602
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1603
+
1604
+ Li et al.
1605
+
1606
+ Figure 8. Flux timeseries of KIC 8478994 before (top) and after (bottom) a whitening filter is applied. The length of the data segment shown in the figures is six days.
1607
+ A single transit is visible in each panel.
1608
+
1609
+ valid data points and outliers, usually resulting in biases in the
1610
+ fitted parameters.
1611
+
1612
+ The robust fitting algorithm, which is optional, works by
1613
+ assigning a weight in the range between 0 and 1 to each data point
1614
+ for the fit. The outliers are assigned weights close to 0 so that the
1615
+ output of the robust fitting algorithm is less sensitive to the outliers
1616
+ in the data. The robust fitting algorithm is executed after the
1617
+ the non-robust LM fit. The weights are
1618
+ convergence of
1619
+ reinitialized and the LM fit is done iteratively. In each iteration,
1620
+ the weight of each data point is calculated from the fit residual of
1621
+ the previous iteration with a bisquare function, so that the data
1622
+ points with larger residuals are assigned smaller weights. The
1623
+ iterative process of weight reassignment and LM fit continues
1624
+ until the fitted parameters converge within a specified tolerance.
1625
+ The effect of robust fit is demonstrated with the following
1626
+ example. The primary and secondary eclipses of an eclipsing
1627
+ binary target (KIC 6960446) are identified as one TCE by the
1628
+ TPS component. Figure 9 shows the folded flux timeseries of
1629
+ the target and the folded model light curve generated with the
1630
+ fitted parameters of the TCE, both in the whitened domain,
1631
+
1632
+ when the robust fit is off (top) and on (bottom), respectively.
1633
+ The secondary eclipse, which has a smaller depth, is located at
1634
+ phase 0 of the plot. A small phase offset is observed in the
1635
+ folded primary and secondary eclipses. In the plot, the flux data
1636
+ points are illustrated as dark green dots when the weights of the
1637
+ robust fit are larger than 0.1, otherwise, illustrated as light blue
1638
+ dots. When the weights of the fit are equally set to 1 for the data
1639
+ points, the fitted model light curve compromises between the
1640
+ primary and secondary eclipses. However, the weights of the fit
1641
+ are calculated iteratively in the robust fitting algorithm. As a
1642
+ result, most data points of the primary eclipses are identified as
1643
+ outliers and assigned weights less than 0.1 at the end of the
1644
+ iterative process. The robust fit algorithm generates unbiased
1645
+ fitted parameters to characterize the secondary eclipses only.
1646
+
1647
+ 6.1.4. Goodness-of-fit Metrics
1648
+ Two goodness-of-fit metrics are calculated when the transit
1649
+ model-fitting algorithm is
1650
+ completed successfully. One
1651
+ includes the χ2 metric and the number of degrees of freedom,
1652
+ the other is the signal-to-noise ratio (S/N) of the fit.
1653
+
1654
+ 12
1655
+
1656
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1657
+
1658
+ Li et al.
1659
+
1660
+ Figure 9. Folded flux timeseries of KIC 6960446 and model light curve generated with fitted parameters of the TCE, both in the whitened domain, when robust fit is
1661
+ disabled (top) and robust fit is enabled (bottom).
1662
+
1663
+ The χ2 metric is determined with Equation (12), and the
1664
+ number of degrees of freedom is determined as the sum of
1665
+ the weights minus the number of fitted parameters. It is noted
1666
+ the weights take values of either 0 or 1 when the robust fit is
1667
+ disabled, as described in Section 6.1.3.
1668
+ The S/N of the fit is determined as:
1669
+
1670
+ /
1671
+ S N
1672
+
1673
+ fit
1674
+
1675
+ = 
1676
+ s
1677
+
1678
+ T^
1679
+ q
1680
+ ( )
1681
+
1682
+ ^
1683
+
1684
+ q
1685
+ W s
1686
+ ( )
1687
+
1688
+ ,
1689
+
1690
+ (
1691
+
1692
+ 14
1693
+
1694
+ )
1695
+
1696
+ where qˆ is the vector of fitted parameters and s q˜( ˆ ) is the
1697
+ whitened model light curve generated with qˆ . W is a diagonal
1698
+ matrix of robust weights as before.
1699
+
1700
+ The χ2 metric and the number of degrees of freedom
1701
+ measure the distance between the flux timeseries and the model
1702
+ light curve in the whitened domain. The S/N shows the
1703
+ strength of the TCE relative to the noise.
1704
+
1705
+ 6.1.5. Uncertainties of Fitted and Derived Parameters
1706
+
1707
+ Let H denote the Jacobian of the model light curve s q˜( ) to
1708
+
1709
+ the vector of fitted parameters q, such that
1710
+
1711
+ Based on the approximation to the Hessian, the covariance
1712
+
1713
+ matrix of the fitted parameters is determined as
1714
+
1715
+ Cov
1716
+
1717
+ ( ˆ )
1718
+ q
1719
+
1720
+ =
1721
+
1722
+ T
1723
+
1724
+ -
1725
+ H WH
1726
+ )
1727
+
1728
+ (
1729
+
1730
+ 1
1731
+
1732
+ (
1733
+
1734
+ s
1735
+ res
1736
+
1737
+ )
1738
+
1739
+ 2
1740
+
1741
+ ,
1742
+
1743
+ (
1744
+
1745
+ 16
1746
+
1747
+ )
1748
+
1749
+ where σres is the root of the mean squared (rms) value of the
1750
+ residuals of the fit. The elements of the Jacobian H are
1751
+ determined numerically.
1752
+
1753
+ Let a and y denote vectors of stellar parameters and derived
1754
+
1755
+ parameters, as defined below:
1756
+
1757
+ a = ⎡⎣
1758
+
1759
+ R g T
1760
+
1761
+ s
1762
+
1763
+ eff
1764
+
1765
+ ⎤⎦
1766
+ T
1767
+
1768
+ and
1769
+
1770
+ y
1771
+
1772
+ =
1773
+
1774
+ ⎡⎣
1775
+
1776
+ R a i d
1777
+
1778
+ p
1779
+
1780
+ tr
1781
+
1782
+ d D T
1783
+
1784
+ in
1785
+
1786
+ eq
1787
+
1788
+ (
1789
+
1790
+ 17
1791
+
1792
+ )
1793
+
1794
+ (
1795
+
1796
+ 18
1797
+
1798
+ )
1799
+
1800
+ T
1801
+
1802
+ ⎤⎦
1803
+
1804
+ ,
1805
+
1806
+ f
1807
+ eff
1808
+
1809
+ where g is the acceleration due to gravity at the surface of the
1810
+ star, determined from the stellar parameter
1811
+ as shown in
1812
+ Equation (4). The uncertainty of g (m s−2) can be determined
1813
+ (log10(cm s−2)) multiplied
1814
+ g
1815
+ from the uncertainty of
1816
+ by g ln 10.
1817
+
1818
+ log
1819
+
1820
+ log
1821
+
1822
+ g
1823
+
1824
+ H
1825
+
1826
+ =
1827
+
1828
+
1829
+
1830
+ ˜( )
1831
+ q
1832
+ s
1833
+
1834
+ q
1835
+
1836
+ .
1837
+
1838
+ (
1839
+
1840
+ 15
1841
+
1842
+ )
1843
+
1844
+ As discussed in Section 4.2, y is a function of q and a. The
1845
+ covariance matrix of y, Cov y(
1846
+ includes the components
1847
+ propagated both from the covariance matrix of q, Cov q( ), and
1848
+
1849
+ ),
1850
+
1851
+ 13
1852
+
1853
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
1854
+
1855
+ Li et al.
1856
+
1857
+ the uncertainties of the elements of a, as shown below:
1858
+
1859
+ Cov
1860
+
1861
+ (
1862
+
1863
+ y
1864
+
1865
+ )
1866
+
1867
+ =
1868
+
1869
+ ⎛
1870
+ ⎜
1871
+ ⎝
1872
+
1873
+ y
1874
+
1875
+
1876
+ q
1877
+
1878
+ ⎞
1879
+ T
1880
+ ⎟
1881
+ ⎠
1882
+
1883
+ Cov
1884
+
1885
+ q
1886
+ ( )
1887
+
1888
+ ⎛
1889
+ ⎜
1890
+ ⎝
1891
+
1892
+ y
1893
+
1894
+
1895
+ q
1896
+
1897
+ ⎞
1898
+ ⎟
1899
+ ⎠
1900
+
1901
+ +
1902
+
1903
+ ⎛
1904
+ ⎜
1905
+ ⎝
1906
+
1907
+
1908
+
1909
+
1910
+ y
1911
+ a
1912
+
1913
+ ⎞
1914
+ T
1915
+ ⎟
1916
+ ⎠
1917
+
1918
+ Cov
1919
+
1920
+ (
1921
+
1922
+ a
1923
+ )
1924
+
1925
+ ⎛
1926
+ ⎜
1927
+ ⎝
1928
+
1929
+
1930
+
1931
+
1932
+ ⎞
1933
+ ⎟
1934
+ ⎠
1935
+
1936
+ y
1937
+ a
1938
+ (
1939
+
1940
+ 19
1941
+
1942
+ )
1943
+
1944
+ where Cov a(
1945
+ ) is a diagonal matrix whose elements are squares
1946
+ of the uncertainties of the corresponding stellar parameters.
1947
+ Note that uncertainties in stellar parameters are provided by the
1948
+ they are assumed to be
1949
+ KIC or overrides to the KIC;
1950
+ independent. y q
1951
+
1952
+
1953
+ are Jacobians, which are
1954
+ described in the Appendix.
1955
+
1956
+
1957
+ ¶ and y a
1958
+
1959
+ The uncertainties of the fitted and derived parameters, the
1960
+ elements of vectors q and y, are determined as the square roots
1961
+ of the diagonal elements of the matrices Cov q( ) and Cov y(
1962
+ ),
1963
+ respectively.
1964
+
1965
+ 6.2. Reduced-parameter Fits
1966
+ Of the five fitted parameters of the geometric transit model
1967
+ defined in Section 4, the impact parameter, b, ranging between
1968
+ 0 and 1, basically describes the slope of the edges of transits.
1969
+ When b is closer to 0, the edges are steeper. Due to the limb-
1970
+ darkening effect of the host star, it is difficult to determine
1971
+ exactly where the transit edges start and end. Therefore, in case
1972
+ of a low S/N for the flux timeseries,
1973
+ there is insufficient
1974
+ information to determine the impact parameter, which leads to
1975
+ large uncertainties in the fitted parameters. When DV is run
1976
+ with different hardware or in different computational environ-
1977
+ ments, the results of the geometric transit model fit may be
1978
+ inconsistent, even with the same code and input data. To
1979
+ reduced-parameter fits are
1980
+ resolve this problem, a set of
1981
+ implemented before the geometric model fitting of all transits
1982
+ and odd–even transits: the impact parameter, b, is set to fixed
1983
+ values 0.1, 0.3, 0.5, 0.7, and 0.9, and only the parameters tepoch,
1984
+ P, Rp/Rs, and a/Rs are allowed to vary. After completion of the
1985
+ reduced-parameter fits,
1986
+ the all-transit fit and the odd–even
1987
+ transit fit follow with initial values set to the fitted parameters
1988
+ of the reduced-parameter fit with the minimum χ2 metric and
1989
+ the corresponding value of the impact parameter.
1990
+ Figure 10 shows the diagnostic plots of
1991
+
1992
+ the reduced-
1993
+ parameter fits of
1994
+ the target star KIC
1995
+ 6541920. As shown in the figure, as the fixed value of b
1996
+ increases from 0.1 to 0.9, the χ2 metric varies less than 0.2% in
1997
+ the reduced-parameter fits. However, Rp/Rs
1998
+ increases by
1999
+ approximately 20% and a/Rs decreases by more than 50%.
2000
+ The results of the reduced-parameter fit with the minimum χ2
2001
+ metric are labeled with red dashed lines in the figure. As
2002
+ illustrated in Figure5 of Section 5, an increase in Rp/Rs leads
2003
+ to an increase in the transit depth, an increase in a/Rs leads to a
2004
+ decrease in the transit duration, and an increase in b results in
2005
+ the decrease in both the transit depth and duration. The
2006
+ observations of Figure 5 are consistent with the systematic
2007
+
2008
+ the sixth TCE of
2009
+
2010
+ variations in Rp/Rs and a/Rs versus b in the reduced-parameter
2011
+ fits shown in Figure 10: when the fixed value of b increases,
2012
+ both the transit depth and duration tend to decrease. Therefore,
2013
+ Rp/Rs increases and a/Rs decreases to compensate for the
2014
+ effect of the increase of b, so that a good fit of the model light
2015
+ curve to the flux timeseries is achieved.
2016
+
2017
+ The plot on the top of Figure 11 shows the light curves
2018
+ generated by the geometric transit signal generator with the
2019
+ fixed values of b and the corresponding sets of fitted parameters
2020
+ tepoch, P, Rp/Rs, and a/Rs of the reduced-parameter fits of the
2021
+ sixth TCE of KIC 6541920. The light curves corresponding to
2022
+ the fixed b values of 0.1, 0.3, 0.5, 0.7, and 0.9 are plotted as
2023
+ blue, red, black, magenta, and green lines, respectively. The
2024
+ plot on the bottom of Figure 11 shows the differences between
2025
+ light curves with fixed b values of 0.3, 0.5, 0.7, and 0.9 and the
2026
+ one with b=0.1. It is observed that the difference in the light
2027
+ curves with different values of b is small; therefore, any small
2028
+ variation in the input flux timeseries may result in a large
2029
+ change in the fitted parameters of Rp/Rs, a/Rs, and b in the all-
2030
+ transit fit and odd–even transit fit.
2031
+
2032
+ 6.3. Odd–Even Transit Fit
2033
+ When the fitting of all transits converges successfully, the
2034
+ same fitting algorithm is executed to fit the odd and even
2035
+ transits to the geometric transit model. The results of the odd–
2036
+ even transit fits are used in the diagnostic tests of the DV
2037
+ component to identify false positives generated by a circular
2038
+ eclipsing binary target or background eclipsing binary.
2039
+
2040
+ The depths of multiple transits of a planet are ideally the
2041
+ same, and the transits of a planet are evenly spaced in time
2042
+ (in the absence of significant
2043
+ transit-timing variations). In
2044
+ contrast, the depths of primary and secondary eclipses of an
2045
+ eclipsing binary system are generally different due to the
2046
+ difference in size and brightness of the two stars. In the odd–
2047
+ even transit fit, two sets of parameters, one set for odd transits
2048
+ and the other set for even transits, are determined through an
2049
+ iterative whitening and model-fitting process described in
2050
+ Section 6.1, and the derived parameters are calculated for
2051
+ each. For each TCE, the transit depths and epochs and the
2052
+ corresponding uncertainties derived from the odd–even transit
2053
+ fit are used in the eclipsing binary discrimination tests to
2054
+ distinguish the flux timeseries of an eclipsing binary system
2055
+ whose primary and secondary eclipses are identified as one
2056
+ TCE in the TPS component. That is, the trial orbital period
2057
+ identified in TPS is half the true orbital period, so that the
2058
+ secondary eclipses are folded on top of the primary eclipses.
2059
+ The details of the eclipsing binary discrimination tests in the
2060
+ DV component are discussed in Twicken et al. (2018).
2061
+
2062
+ Figure 12 shows the folded unwhitened flux timeseries of the
2063
+ odd and even transits of the eclipsing binary target KIC
2064
+ 6960446. Figure 13 shows the folded whitened flux timeseries
2065
+
2066
+ 14
2067
+
2068
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2069
+
2070
+ Li et al.
2071
+
2072
+ Figure 10. Reduced-parameter fits of the sixth TCE of KIC 6541920: χ2 metric (top), fitted parameters Rp/Rs (middle), and a/Rs (bottom) vs. impact parameter, b.
2073
+
2074
+ of the odd and even transits of the same target and the folded
2075
+ whitened model light curves generated with fitted parameters of
2076
+ the odd and even transits, respectively. As shown in the figures,
2077
+ the primary and secondary eclipses are identified as one TCE
2078
+ by the TPS component, the fits of odd and even transits, which
2079
+ are actually primary and secondary eclipses, demonstrate that
2080
+ the derived transit depths of odd and even transits are different
2081
+ by approximately 15% and that the transit epoch time of the
2082
+ even transits has a small offset of approximately one hour
2083
+ relative to that of the odd transits.
2084
+
2085
+ 6.4. Outputs of Geometric Transit Model Fits
2086
+ When a TCE is identified in the multiple-planet search, as
2087
+ described later in Section 8, a simple check is implemented
2088
+ before fitting the TCE. When the eclipsing depth is more than
2089
+ 250,000 ppm,
2090
+ the TCE is labeled as a suspected eclipsing
2091
+ binary and geometric transit model fitting is not performed.
2092
+
2093
+ When the geometric transit model-fitting algorithm is
2094
+ completed successfully, the fitted parameters and uncertainties,
2095
+ the derived parameters and uncertainties, and the goodness-of-
2096
+ fit metrics, etc. are saved in the DV output structure. In
2097
+
2098
+ 15
2099
+
2100
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2101
+
2102
+ Li et al.
2103
+
2104
+ Figure 11. Top plot shows light curves generated with the geometric transit signal generator with the fixed b values of 0.1, 0.3, 0.5, 0.7, and 0.9 and the corresponding
2105
+ sets of fitted parameters tepoch, P, Rp/Rs, and a/Rs of the reduced-parameter fits of the sixth TCE of KIC 6541920. The bottom plot shows the differences between light
2106
+ curves with fixed b values of 0.3, 0.5, 0.7, and 0.9, and one with b=0.1.
2107
+
2108
+ addition, a set of diagnostic figures are generated by the
2109
+ geometric transit model-fitting algorithm. The diagnostic
2110
+ figures are included in the DV report produced for each target
2111
+ with at least one TCE (Twicken et al. 2016) and archived at the
2112
+ (Akeson et al. 2013). As
2113
+ Exoplanet Archive at NExScI
2114
+ examples, the diagnostic plots of the all-transit fit of the sixth
2115
+ TCE of the target star KIC 6541920 are shown in Figures 14,
2116
+ 15, and 16.
2117
+
2118
+ The plot on the top of Figure 14 shows the detrended, folded
2119
+ unwhitened flux timeseries of all transits of the TCE, and the
2120
+ plot on the bottom of Figure 14 shows the corresponding
2121
+ folded whitened flux timeseries in the same phase range. It is
2122
+ noted that the vertical scales of the two plots in Figure 14 are
2123
+
2124
+ different: the unwhitened flux on the top is dimensionless while
2125
+ the whitened flux on the bottom is in units of the standard
2126
+ deviation of the unwhitened flux. The transit depth derived
2127
+ from the all-transit fit is illustrated with a horizontal red line in
2128
+ the plot on the top. In the plot on the bottom, the folded
2129
+ whitened light curve is illustrated in red, which is generated by
2130
+ the geometric transit signal generator with the fitted parameters
2131
+ derived from the robust fit to all transits. The flux data whose
2132
+ robust weights are larger than 0.1 in the all-transit fit are plotted
2133
+ as dark green dots, otherwise, in light blue dots. The residuals
2134
+ of the fit, determined as the difference of the binned average
2135
+ values of the whitened flux and the whitened model light curve,
2136
+ are plotted as green dots. The same residuals, offset by 180° in
2137
+
2138
+ 16
2139
+
2140
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2141
+
2142
+ Li et al.
2143
+
2144
+ Figure 12. Folded unwhitened flux timeseries of the odd (left) and even (right) transits of KIC 6960446.
2145
+
2146
+ Figure 13. Folded flux timeseries and model light curves, both in whitened domain, of the odd (top) and even (bottom) transits of KIC 6960446.
2147
+
2148
+ phase, are plotted as magenta dots, to aid in the detection of a
2149
+ secondary eclipse. Figure 15 shows the folded weights of the
2150
+ robust fit of the all-transit fit of the sixth TCE of the target star
2151
+ KIC 6541920, in the same phase range as Figure 14.
2152
+
2153
+ Figure 16 shows the detrended, folded unwhitened flux
2154
+ timeseries of the transits of the sixth TCE of the target
2155
+
2156
+ star KIC 6541920 by quarter and season, as well as the
2157
+ corresponding folded unwhitened model
2158
+ light curves of
2159
+ the all-transit fit. The folded transits from the same year of
2160
+ the Kepler mission are plotted in the same row, and the
2161
+ folded transits in the same season are plotted in the
2162
+ same column. For example, the folded transits in Q4 are
2163
+
2164
+ 17
2165
+
2166
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2167
+
2168
+ Li et al.
2169
+
2170
+ Figure 14. Folded flux timeseries and model light curve of the all-transit fit of the sixth TCE of KIC 6541920: unwhitened flux (top), and whitened flux and whitened
2171
+ model light curve (bottom).
2172
+
2173
+ Figure 15. Folded robust weights of the all-transit fit of the sixth TCE of KIC 6541920.
2174
+
2175
+ shown in the upper right corner of the figure. The folded
2176
+ transits of the ���rst year, including Q1, Q2, Q3, and Q4, are
2177
+ shown in the upper left corner, and the folded transits
2178
+ in Season 2, including Q4, Q8, Q12, and Q16, are shown in
2179
+ the lower right corner. At the lower left corner, the folded
2180
+ transits in all 17 quarters of the Kepler science data are
2181
+ illustrated.
2182
+
2183
+ For the odd–even transit fit, as illustrated in Figures 12 and 13,
2184
+ the plots of folded unwhitened flux timeseries of odd and even
2185
+ transits are placed horizontally, and the plots of folded flux
2186
+ timeseries and folded model light curves, both whitened, are
2187
+ placed vertically, so that the difference in the derived depths and
2188
+ the offset in the fitted transit epoch times of the odd and even
2189
+ transits can be easily observed.
2190
+
2191
+ 18
2192
+
2193
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2194
+
2195
+ Li et al.
2196
+
2197
+ Figure 16. Folded flux timeseries centered on the transit events and folded model light curve of the all-transit fit, both unwhitened, of the sixth TCE of KIC 6541920
2198
+ by quarter and season.
2199
+
2200
+ For the reduced-parameter fits, a set of diagnostic plots, the
2201
+ same as those of the all-transit fit, are generated for each fit. In
2202
+ addition, as illustrated in Figure 10, several diagnostic plots are
2203
+ generated to illustrate the variations of the χ2 metric and fitted
2204
+ parameters versus the fixed value of the impact parameter. The
2205
+ fit with the minimum χ2 metric is labeled with red dash lines on
2206
+ these figures.
2207
+
2208
+ 6.5. Alerts of Failed Fits
2209
+ When the geometric transit model fit fails, an alert
2210
+ is
2211
+ generated indicating the nature of the failure and where it
2212
+ occurs. These alerts are included in the Appendix of the DV
2213
+
2214
+ report. Table 3 lists the top five alerts of the failed all-transit fits
2215
+ in the DR25 DV processing.
2216
+
2217
+ As shown the the table, the most common failure of the all-
2218
+ transit fits is that the time used by the fitting algorithm goes
2219
+ beyond the preset limit and the fit is stopped during the call of
2220
+ the function “model_function.” This usually happens when an
2221
+ anomalously noisy flux timeseries is fitted with a transit model;
2222
+ the criterion of convergence can never be met and the algorithm
2223
+ goes into an infinite loop.
2224
+
2225
+ In the fitting algorithm, several check points are set to verify
2226
+ the validity of the fit results, or else the fit results are labeled as
2227
+ invalid. For example, the fitted parameter of the transit epoch
2228
+ time tepoch should fall in a range centered on the corresponding
2229
+
2230
+ 19
2231
+
2232
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2233
+
2234
+ Li et al.
2235
+
2236
+ Figure 17. Folded flux timeseries and folded model light curve of the trapezoidal model fit, both unwhitened, of the sixth TCE of KIC 6541920.
2237
+
2238
+ Table 3
2239
+ Top Five Alerts of Failed All-transit Fits in DR25 DV Run
2240
+
2241
+ Index
2242
+
2243
+ 1
2244
+ 2
2245
+
2246
+ 3
2247
+
2248
+ 4
2249
+ 5
2250
+
2251
+ Alert Type
2252
+ dv:modelFunction:fitTimeLimitExceeded
2253
+ dv:fitTransit:
2254
+
2255
+ transitEpochBkjdBigDifferenceFromTceValue
2256
+
2257
+ dv:fillPlanetResults:
2258
+
2259
+ transitDurationSmallerThanLowerBound
2260
+
2261
+ dv:computeLargeBodyTransitLightCurve:takingTooLong
2262
+ dv:transitFitClass:insufficientTransitsToFit
2263
+
2264
+ 1,012
2265
+ 592
2266
+
2267
+ 262
2268
+
2269
+ 45
2270
+ 41
2271
+
2272
+ TCE value given by the TPS component. Furthermore, the
2273
+ derived transit duration cannot be smaller than the duration of a
2274
+ LC interval (29.4 min). As shown in items 2 and 3 of Table 3,
2275
+ the alerts of invalid fit results are generated during the call of
2276
+ the functions “fit_transit” and “fill_planet_results.”
2277
+
2278
+ The fourth alert of Table 3 is generated when the time used in
2279
+ the iterative numerical
2280
+ integration algorithm, as described in
2281
+ Section 5, exceeds the preset limit in the call of the function to
2282
+ compute the transit light curve when the small-body approx-
2283
+ imation is not applicable. The fifth alert occurs during the call of
2284
+ the function “transitFitClass” when too many flux data points are
2285
+ gapped and the number of remaining transits is less than 2 in the
2286
+ all-transit fit; as a result,
2287
+ information to
2288
+ determine reliable parameters of the transit model.
2289
+
2290
+ there is insufficient
2291
+
2292
+ 7. Trapezoidal Model-fitting Algorithm
2293
+ As an optional configuration of the transit model fitting in
2294
+ the DV component, the light curve of the target for which a
2295
+ TCE is generated can also be fitted by a trapezoidal model. The
2296
+ trapezoidal model
2297
+ the basic
2298
+ characteristics of the transits, and may converge to a successful
2299
+ fit when the limb-darkened transit model fit fails. In these cases,
2300
+ the trapezoidal model fit parameters can be used to support
2301
+
2302
+ is a simple description of
2303
+
2304
+ 20
2305
+
2306
+ subsequent DV diagnostic tests, which otherwise could not be
2307
+ performed (Twicken et al. 2018).
2308
+
2309
+ The trapezoidal model includes the following four fitted
2310
+
2311
+ Number
2312
+
2313
+ parameters:
2314
+
2315
+ (cid:129) Transit epoch time tepoch (BKJD): same as the fitted
2316
+ parameter of the geometric model defined in Section 4.1;
2317
+ (cid:129) Transit depth D (ppm): same as the derived parameter of
2318
+
2319
+ the geometric model defined in Section 4.2;
2320
+
2321
+ (cid:129) Transit mid-duration dmid (hours): the duration of transit at
2322
+ half of the transit depth, as illustrated in Figure 3; and
2323
+ (cid:129) Ratio of ingress time to mid-duration din/dmid (dimension-
2324
+ less): the transit ingress time din is same as the derived
2325
+ parameter defined in Section 4.2, but this is the ratio of the
2326
+ ingress time to mid-duration.
2327
+
2328
+ The orbital period, P, (days) is set to the corresponding TCE
2329
+ parameter value provided by the TPS component in the transit
2330
+ signal generator with the trapezoidal model.
2331
+
2332
+ An alternative detrending algorithm based on the nonpara-
2333
+ metric penalized least-squares method from Garcia (2010) is
2334
+ applied to the PDC light curve prior to the trapezoidal model
2335
+ fit. The algorithm allows for missing data via weight assign-
2336
+ ment and solves for the free parameter controlling the amount
2337
+ of smoothing using a generalized cross validation method. To
2338
+ prevent suppression of the transit signal we treat data in transit
2339
+ according to the TCE ephemeris and transit duration as missing
2340
+ with a weight of zero. Each quarterly PDC light curve is
2341
+ detrended independently. When a high frequency (similar or
2342
+ shorter timescale than the transit signal) astrophysical signal is
2343
+ present in a light curve, the automated method for determining
2344
+ the smoothing parameter results in unwanted suppression of the
2345
+ transit signal. To prevent over-smoothing,
2346
+ the smoothing
2347
+ parameter is determined on a light curve with a low-pass filter
2348
+ applied. The low-pass filtered light curve is generated by
2349
+ subtraction of a high-pass (simple two-point difference) filtered
2350
+ version of the light curve. The adopted detrending model,
2351
+
2352
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2353
+
2354
+ Li et al.
2355
+
2356
+ Figure 18. Light curve of KIC 6541920 from Q1 to Q4 and transits of six TCEs.
2357
+
2358
+ which results from using the smoothing parameter estimated
2359
+ from the low-pass filtered version of the light curve, is used in
2360
+ normalization of the PDC light curve.
2361
+
2362
+ The trapezoidal model-fitting algorithm is implemented with
2363
+ 10 repeated LM fits. For each fit, the initial value of the fitted
2364
+ parameter is set randomly with a uniform distribution in a pre-
2365
+ determined range. The outputs of the trapezoidal model-fitting
2366
+ algorithm are determined as those of the LM fit with the
2367
+ minimum χ2 metric.
2368
+
2369
+ Figure 17 shows a diagnostic plot generated in the trapezoidal
2370
+ model fit of the sixth TCE of KIC 6541920. Only the flux data
2371
+ whose timestamps fall in the time ranges of 8 times the transit
2372
+ duration (one of the TCE parameters generated by the TPS
2373
+ component) and centered at the transit center time are employed in
2374
+ the trapezoidal model fit. The flux data points within this range
2375
+ used in the fit are plotted as dark green dots in the figure,
2376
+ otherwise, in light blue dots. The folded light curve generated by
2377
+ the trapezoidal model with the fitted parameters is plotted as red
2378
+ lines and the residual of the fit is offset vertically for clarity and
2379
+ plotted as green dots. Because the whitening filter, described in
2380
+ Section 6.1.1, is not used in the trapezoidal model-fitting algorithm,
2381
+ all the data shown in Figure 17 are in the unwhitened domain.
2382
+
2383
+ 21
2384
+
2385
+ Compared to the plot on the bottom of Figure14 of
2386
+ Section 6.4, the bottom of the transit is flat in the model light
2387
+ curve shown in Figure 17 because the limb-darkening effect is
2388
+ not included in the trapezoidal transit model.
2389
+
2390
+ The trapezoidal model fit provides a quick assessment of the
2391
+ transit signal. The fitted trapezoidal transit model is used in
2392
+ the diagnostic tests of the DV component when the fit with the
2393
+ geometric transit model fails or when the fit is not performed,
2394
+ such as for suspected eclipsing binaries.
2395
+
2396
+ time of
2397
+
2398
+ 8. Multiple-planet Search
2399
+ After the fitting process has completed,
2400
+
2401
+ the data points
2402
+ within 1.5 times the transit duration from the central time of the
2403
+ nearest transit are removed, where the transit duration and the
2404
+ transits are determined from the fitted
2405
+ central
2406
+ parameters of the all-transit fit. So the signature of the known
2407
+ TCE is removed, and the residual flux is subjected to a search
2408
+ for additional planets by calling TPS in the DV component.
2409
+ The transit model-fitting algorithms, including the reduced-
2410
+ parameter fits, all-transit fit, odd–even transit fit, and the
2411
+ trapezoidal model fit, are applied again if an additional TCE is
2412
+ generated. The search for additional planets concludes when no
2413
+
2414
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2415
+
2416
+ Li et al.
2417
+
2418
+ Figure 19. Phased flux timeseries of KIC 6541920 with the fitted parameters tepoch and P of the fifth (top) and sixth (bottom) TCEs, respectively.
2419
+
2420
+ additional TCEs are produced or an iteration limit is reached, as
2421
+ shown in the flowchart of Figure 1.
2422
+
2423
+ Figure 18 shows the light curve of KIC 6541920 (Kepler-11)
2424
+ from Q1 to Q4. The quarterly segments are offset vertically for
2425
+ clarity. The transits of six TCEs are labeled with different
2426
+ colors and symbols in the figure. The first TCE, labeled with
2427
+ is identified by the TPS component and the
2428
+ red circles,
2429
+ corresponding parameters to characterize the TCE are provided
2430
+ to DV. The remaining five TCEs are identified in the multiple-
2431
+ planet search by calling TPS directly in the DV component.
2432
+
2433
+ Figure 19 shows the folded flux timeseries of KIC 6541920
2434
+ in the unwhitened domain, phased with the fitted parameters
2435
+ tepoch and P of the fifth and sixth TCEs, respectively. The
2436
+ binned average values of the folded flux and the folded model
2437
+ light curve are plotted as blue and red dots, respectively. The
2438
+ triangles in different colors show the location of the transits of
2439
+ all six TCEs in the phased flux timeseries.
2440
+
2441
+ 9. Performance of Transit Model-fitting and
2442
+ Multiple-planet Search
2443
+
2444
+ The 17 quarters of primary mission science data, collected by
2445
+ the Kepler spacecraft from May 13, 2009 to April 8, 2013, were
2446
+ processed by the SOC 9.3 codebase of the Kepler Data Processing
2447
+ Pipeline in January 2016. 17,230 target stars, which generated
2448
+ TCEs in the TPS component, were processed successfully by the
2449
+
2450
+ 22
2451
+
2452
+ DV component. This pipeline run is referred to as DR25, and the
2453
+ TCE population was described in Twicken et al. (2016).
2454
+
2455
+ Among a total of 34,032 TCEs generated in the TPS component
2456
+ and in the multiple-planet search of the DV component, 239
2457
+ (0.7%) TCEs were labeled as suspected eclipsing binaries, 2,062
2458
+ (6.1%) TCEs failed in the all-transit fit, and 31,731 (93.2%) TCEs
2459
+ completed the all-transit fit successfully. Out of 31,731 TCEs with
2460
+ successful all-transit fits, 2,620 (8.3%) TCEs failed in the odd–
2461
+ even transit fit, and 29,111 (91.7%) TCEs completed the odd–even
2462
+ transit fit successfully. 33,125 (97.3%) out of 34,032 TCEs
2463
+ completed the trapezoidal model fit successfully.
2464
+
2465
+ Figure 20 compares the orbital period of the DV all-transit fit
2466
+ and the corresponding KOI parameter produced independently
2467
+ (Rowe et al. 2014). The plot on the left shows all orbital
2468
+ periods in the comparison and the plot on the right shows the
2469
+ orbital periods ranging from 0 to 20 days only. The diagonal
2470
+ green line shows where the DV fitted orbital period value is
2471
+ equal to the KOI parameter value; the other four green lines
2472
+ indicate that the two period values differ by a factor of 1/3, 1/
2473
+ 2, 2, and 3, respectively. It is observed in Figure 20 that the
2474
+ orbital periods of some TCEs identified in TPS and DV are
2475
+ double or half of the corresponding KOI values.
2476
+
2477
+ Figure 21 compares the transit depth derived from the
2478
+ DV all-transit fit and the corresponding KOI parameter.
2479
+ Similar to Figure 20, the plot on the left shows all-transit
2480
+
2481
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2482
+
2483
+ Li et al.
2484
+
2485
+ Figure 20. Comparison of DV Fitted parameters and KOI parameters: all orbital periods (left) and orbital periods ranging from 0 to 20 days (right).
2486
+
2487
+ Figure 21. Comparison of DV Fitted parameters and KOI parameters: all-transit depths (left) and transit depths ranging from 0 to 500 ppm (right).
2488
+
2489
+ depths in the comparison and the plot on the right shows the
2490
+ transit depths ranging from 0 to 500 ppm only. The diagonal
2491
+ green line shows where the DV fitted transit depth value
2492
+ is equal
2493
+ is observed that
2494
+ the KOI values of the transit depth are larger than the
2495
+
2496
+ to the KOI parameter value. It
2497
+
2498
+ corresponding DV fitted values for many TCEs. Investiga-
2499
+ tions show some short-period transit signals are degraded
2500
+ in the light curve preprocessing procedure of harmonic
2501
+ removal when the orbital period is small (Christiansen et al.
2502
+ 2013, 2015).
2503
+
2504
+ 23
2505
+
2506
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2507
+
2508
+ Li et al.
2509
+
2510
+ Figure 22. Diagnostic plots of χ2 vs. b of the reduced-parameter fits of the 1st TCE of KIC 6541920, generated by the SOC 9.3 codebase in 2016 January (top) and
2511
+ the modified SOC 9.3 codebase in 2016 August 2 (bottom), respectively. As shown in the plot on the top, due to a software defect introduced into the 9.3 codebase, the
2512
+ χ2 metric of the reduced-parameter fit systematically decreases as the fixed value of the impact parameter b increases. In the plot on the bottom, there is no systematic
2513
+ decrease of the χ2 metric as b increases.
2514
+
2515
+ A software defect introduced into the SOC 9.3 code for the
2516
+ reduced-parameter fits came to light after the DR25 run. As
2517
+ discussed in Section 6.1, only the data points within the range
2518
+ of the transit and a buffer on each side of the transit are
2519
+ employed in the weighted nonlinear least-squares fitting. The
2520
+ weights are assigned 1 and 0, respectively, depending on
2521
+ whether the data points are used in the fitting or not. As shown
2522
+ in Equation (12), the χ2 metric is related to how many data
2523
+ points are used in the fit: the more data points used in the fit, the
2524
+ larger the χ2 metric. In the SOC 9.3 codebase, the data points
2525
+ employed in the reduced-parameter fits are related to the fixed
2526
+ value of the impact parameter b. As a result, the calculated χ2
2527
+ metric is improperly related to the value of b: the closer b is to
2528
+ 1, the smaller the χ2 metric. The software defect was corrected
2529
+
2530
+ in a modified SOC 9.3 codebase, which was used in a
2531
+ supplemental DV run in August 2016. Figure 22 shows the
2532
+ diagnostic plots of the χ2 metric versus b of the reduced-
2533
+ parameter fits of the 1st TCE of KIC 6541920 (the planet
2534
+ Kepler-11e), which were generated by the SOC 9.3 codebase in
2535
+ January 2016 and the modified SOC 9.3 codebase in August
2536
+ 2016, respectively. As shown in the plot on the top of
2537
+ Figure 22, due to the software defect, the χ2 metric system-
2538
+ atically decreases as b increases so the result of the reduced-
2539
+ parameter fit with the fixed value of b=0.9 is always selected
2540
+ to seed the all-transit fit. The same was true for all TCEs in the
2541
+ DR25 DV run. In the plot on the bottom of Figure 22, there is
2542
+ no systematic decrease of the χ2 metric as b increases, and
2543
+ b=0.5 is selected to seed the all-transit fit.
2544
+
2545
+ 24
2546
+
2547
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2548
+
2549
+ Li et al.
2550
+
2551
+ Figure 23. Distribution of the fitted parameter b of the all-transit fits of a set of 16,514 TCEs, generated by the SOC 9.3 codebase in 2016 January (top) and the
2552
+ modified SOC 9.3 codebase in 2016 August (bottom), respectively. As shown in the plot on the top, the distribution of the fitted parameter b is biased toward b=0.9
2553
+ in the outputs of the all-transit fit of the SOC 9.3 codebase. In the plot on the bottom, there is no bias toward b=0.9 in the distribution of the fitted parameter b in the
2554
+ outputs of the all-transit fit of the modified SOC 9.3 codebase.
2555
+
2556
+ As shown in Section 6.2, flux timeseries with low S/N
2557
+ including those with transiting planet signatures of small
2558
+ planets (relative to the size of their host stars) may be well fitted
2559
+ over a wide range of impact parameter values. Figure 23 shows
2560
+ the distributions of the fitted parameter b in the all-transit fits of
2561
+ a set of 16,514 TCEs, generated in the DR25 run with the SOC
2562
+ 9.3 codebases in 2016 January and in the supplemental DV run
2563
+ with the modified SOC 9.3 codebase in 2016 August,
2564
+ respectively. The 16,514 TCEs were selected from the 1st
2565
+ the targets, which completed the all-transit fits
2566
+ TCEs of
2567
+ the fitted
2568
+ successfully in both runs. The distribution of
2569
+ parameter b is biased toward the initial seed value of b=0.9
2570
+ in the outputs of the all-transit fits with the SOC 9.3 codebase,
2571
+
2572
+ as shown in the plot on the top of Figure 23. In the plot on the
2573
+ bottom of Figure 23, there is no bias toward b=0.9 in the
2574
+ distribution of the fitted parameter b in the all-transit fits with
2575
+ the modified SOC 9.3 codebase. Figure 24 shows the
2576
+ distributions of the fitted parameter b in the all-transit fits of
2577
+ a set of 1,292 TCEs in both DV runs. The set of 1,292 TCEs, a
2578
+ subset of the 16,514 TCEs, was selected as the fitted parameter
2579
+ Rp/Rs was larger than 0.1 in the supplemental DV run in 2016
2580
+ August. It is observed that the convergence of the all-transit fit
2581
+ is essentially independent of the initial seed value of the impact
2582
+ parameter b for large planets.
2583
+
2584
+ As discussed by Twicken et al. (2016), transiting planets
2585
+ with a high impact parameter must be larger than those with a
2586
+
2587
+ 25
2588
+
2589
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2590
+
2591
+ Li et al.
2592
+
2593
+ Figure 24. Distribution of the fitted parameter b of the all-transit fits of a set of 1,292 TCEs, generated by the SOC 9.3 codebase in 2016 January (top) and the
2594
+ modified SOC 9.3 codebase in 2016 August (bottom), respectively. The set of 1,292 TCEs, a subset of the 16,514 TCEs, was selected as the fitted parameter Rp/Rs
2595
+ was larger than 0.1 in the supplemental DV run in 2016 August. It is observed that the convergence of the all-transit fit is essentially independent of the initial seed
2596
+ value of the impact parameter b for large planets.
2597
+
2598
+ lower impact parameter for given transit depths on the same
2599
+ host stars because of the limb-darkening effect. It is noted that
2600
+ all planetary candidates in the DR25 Kepler Mission catalog by
2601
+ Thompson et al. (2018) were modeled independently by the
2602
+ TCE Review Team (TCERT), so the bias discussed here relates
2603
+ only to TCE products of the SOC 9.3 DR25 of the Kepler
2604
+ Science Data Processing Pipeline at
2605
+ the NASA Exoplanet
2606
+ Archive.
2607
+
2608
+ 10. Conclusions
2609
+ We have presented the transit model-fitting and multiple-
2610
+ planet search algorithm of the Data Validation component of
2611
+ the Kepler Science Data Processing Pipeline. The performance
2612
+ of the algorithm is demonstrated by the results of processing 17
2613
+ quarters of Kepler science data using SOC 9.3 codebase of the
2614
+ Kepler Science Data Processing Pipeline in January 2016
2615
+
2616
+ (DR25). The results of the transit model fitting of the TCEs
2617
+ identified by the pipeline are accessible by the science
2618
+ the NASA Exoplanet Archive. The Kepler
2619
+ community at
2620
+ SOC 9.3 codebase is also available to the general public
2621
+ through GitHub. A software defect that biased the seeding of
2622
+ the limb-darkened model fits and ultimately the model fit
2623
+ results for small planets was corrected in a modified SOC 9.3
2624
+ codebase, which was implemented in a supplemental DV run
2625
+ after DR25.
2626
+
2627
+ Funding for the KeplerMission was provided by the NASA
2628
+ Science Mission Directorate. The data validation products were
2629
+ generated by the Kepler Science Data Processing Pipeline
2630
+ through the efforts of the Kepler Science Operations Center and
2631
+ Science Office at NASA Ames Research Center.
2632
+
2633
+ Facility: Kepler.
2634
+
2635
+ 26
2636
+
2637
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
2638
+
2639
+ Appendix
2640
+ Jacobians in Section 6.1.5
2641
+
2642
+
2643
+ a
2644
+
2645
+ P
2646
+
2647
+ =
2648
+
2649
+ 2
2650
+ 3
2651
+
2652
+ a
2653
+ P
2654
+
2655
+ ,
2656
+
2657
+ Li et al.
2658
+
2659
+ (
2660
+
2661
+ 23
2662
+
2663
+ )
2664
+
2665
+ The Jacobians y q
2666
+
2667
+
2668
+ ¶ and y a
2669
+
2670
+ ¶ in Section 6.1.5 have the
2671
+
2672
+
2673
+
2674
+ following forms:
2675
+
2676
+
2677
+ i
2678
+ a R
2679
+ s
2680
+
2681
+ )
2682
+
2683
+
2684
+
2685
+ (
2686
+
2687
+ =
2688
+
2689
+ 180
2690
+ p
2691
+
2692
+ b
2693
+ a R
2694
+ s
2695
+
2696
+ (
2697
+
2698
+ )
2699
+
2700
+ 1
2701
+
2702
+ (
2703
+
2704
+ a R
2705
+ s
2706
+
2707
+ )
2708
+
2709
+ 2
2710
+
2711
+ -
2712
+
2713
+ 2
2714
+
2715
+ b
2716
+
2717
+ ,
2718
+
2719
+ (
2720
+
2721
+ 24
2722
+
2723
+ )
2724
+
2725
+ y
2726
+
2727
+
2728
+ q
2729
+
2730
+ =
2731
+
2732
+ ⎡
2733
+ ⎢
2734
+ ⎢
2735
+ ⎢
2736
+ ⎢
2737
+ ⎢
2738
+ ⎢
2739
+ ⎢
2740
+ ⎢
2741
+ ⎢
2742
+ ⎢
2743
+ ⎢
2744
+ ⎢
2745
+ ⎢
2746
+ ⎢
2747
+ ⎢
2748
+ ⎢
2749
+ ⎢
2750
+ ⎢
2751
+ ⎢
2752
+ ⎢
2753
+ ⎢
2754
+ ⎢
2755
+ ⎣
2756
+
2757
+ 0
2758
+
2759
+ 0
2760
+
2761
+ 0
2762
+
2763
+ 0
2764
+
2765
+ 0
2766
+
2767
+ 0
2768
+
2769
+ 0
2770
+
2771
+ 0
2772
+
2773
+ 0
2774
+
2775
+
2776
+ a
2777
+
2778
+ P
2779
+
2780
+ 0
2781
+
2782
+
2783
+ d
2784
+ tr
2785
+
2786
+ P
2787
+
2788
+ d
2789
+ in
2790
+
2791
+ P
2792
+
2793
+ 0
2794
+
2795
+
2796
+
2797
+ T
2798
+ eq
2799
+
2800
+ P
2801
+
2802
+ f
2803
+ eff
2804
+ P
2805
+
2806
+
2807
+
2808
+ )
2809
+
2810
+ 0
2811
+
2812
+ 0
2813
+
2814
+
2815
+ i
2816
+ a R
2817
+ (
2818
+ s
2819
+
2820
+ d
2821
+ tr
2822
+ a R
2823
+ (
2824
+ s
2825
+
2826
+ d
2827
+ in
2828
+ a R
2829
+ (
2830
+ s
2831
+
2832
+ D
2833
+ a R
2834
+ s
2835
+
2836
+ (
2837
+
2838
+
2839
+
2840
+
2841
+
2842
+
2843
+
2844
+
2845
+
2846
+ )
2847
+
2848
+ )
2849
+
2850
+ )
2851
+
2852
+ )
2853
+
2854
+
2855
+
2856
+ R
2857
+
2858
+ p
2859
+ R R
2860
+ s
2861
+ p
2862
+
2863
+
2864
+
2865
+ (
2866
+
2867
+ 0
2868
+
2869
+ 0
2870
+
2871
+
2872
+ d
2873
+ tr
2874
+ R R
2875
+ s
2876
+ p
2877
+
2878
+ d
2879
+ in
2880
+ R R
2881
+ s
2882
+ p
2883
+
2884
+ D
2885
+ R R
2886
+ s
2887
+ p
2888
+
2889
+ )
2890
+
2891
+ )
2892
+
2893
+ )
2894
+
2895
+
2896
+
2897
+ (
2898
+
2899
+
2900
+
2901
+ (
2902
+
2903
+
2904
+
2905
+ (
2906
+
2907
+ 0
2908
+
2909
+ 0
2910
+
2911
+ ⎤
2912
+ ⎥
2913
+ ⎥
2914
+ ⎥
2915
+ ⎥
2916
+ ⎥
2917
+ ⎥
2918
+ ⎥
2919
+ ⎥
2920
+ ⎥
2921
+ ⎥
2922
+ ⎥
2923
+ ⎥
2924
+ ⎥
2925
+ ⎥
2926
+ ⎥
2927
+ ⎥
2928
+ ⎥
2929
+ ⎥
2930
+ ⎥
2931
+ ⎥
2932
+ ⎥
2933
+ ⎥
2934
+ ⎦
2935
+
2936
+ 0
2937
+
2938
+ 0
2939
+
2940
+
2941
+ i
2942
+
2943
+ b
2944
+
2945
+ d
2946
+ tr
2947
+
2948
+ b
2949
+
2950
+ d
2951
+ in
2952
+
2953
+ b
2954
+
2955
+ D
2956
+
2957
+ b
2958
+
2959
+ 0
2960
+
2961
+ 0
2962
+
2963
+ 0
2964
+
2965
+ 0
2966
+
2967
+ 0
2968
+
2969
+ 0
2970
+
2971
+ 0
2972
+ 0
2973
+ 0
2974
+ 0
2975
+ T
2976
+ eq
2977
+
2978
+
2979
+
2980
+
2981
+
2982
+
2983
+
2984
+
2985
+ T
2986
+ eff
2987
+ f
2988
+ eff
2989
+ T
2990
+ eff
2991
+
2992
+ ⎤
2993
+ ⎥
2994
+ ⎥
2995
+ ⎥
2996
+ ⎥
2997
+ ⎥
2998
+ ⎥
2999
+ ⎥
3000
+ ⎥
3001
+ ⎥
3002
+ ⎥
3003
+ ⎥
3004
+ ⎥
3005
+ ⎥
3006
+ ⎥
3007
+ ⎥
3008
+ ⎦
3009
+
3010
+ .
3011
+
3012
+
3013
+
3014
+
3015
+ y
3016
+ a
3017
+
3018
+ =
3019
+
3020
+ ⎡
3021
+ ⎢
3022
+ ⎢
3023
+ ⎢
3024
+ ⎢
3025
+ ⎢
3026
+ ⎢
3027
+ ⎢
3028
+ ⎢
3029
+ ⎢
3030
+ ⎢
3031
+ ⎢
3032
+ ⎢
3033
+ ⎢
3034
+ ⎢
3035
+ ⎢
3036
+ ⎣
3037
+
3038
+
3039
+
3040
+ R
3041
+
3042
+ p
3043
+
3044
+ R
3045
+ s
3046
+
3047
+ a
3048
+
3049
+ R
3050
+ s
3051
+ 0
3052
+ 0
3053
+ 0
3054
+ 0
3055
+ T
3056
+ eq
3057
+
3058
+ R
3059
+ s
3060
+
3061
+ f
3062
+ eff
3063
+ R
3064
+ s
3065
+
3066
+
3067
+
3068
+
3069
+
3070
+ 0
3071
+
3072
+
3073
+ a
3074
+
3075
+ g
3076
+ 0
3077
+ 0
3078
+ 0
3079
+ 0
3080
+ T
3081
+ eq
3082
+
3083
+ g
3084
+
3085
+ f
3086
+ eff
3087
+
3088
+ g
3089
+
3090
+
3091
+
3092
+
3093
+ i
3094
+
3095
+ b
3096
+
3097
+ = -
3098
+
3099
+ 180
3100
+ p
3101
+
3102
+ 1
3103
+
3104
+ (
3105
+
3106
+ a R
3107
+ s
3108
+
3109
+ )
3110
+
3111
+ 2
3112
+
3113
+ -
3114
+
3115
+ 2
3116
+
3117
+ b
3118
+
3119
+ ,
3120
+
3121
+
3122
+ d
3123
+ tr
3124
+ R R
3125
+ s
3126
+ p
3127
+
3128
+ )
3129
+
3130
+
3131
+
3132
+ (
3133
+
3134
+ =
3135
+
3136
+ P
3137
+
3138
+ 24
3139
+ p
3140
+
3141
+
3142
+ d
3143
+ tr
3144
+
3145
+ P
3146
+
3147
+ =
3148
+
3149
+ d
3150
+ tr
3151
+ P
3152
+
3153
+ ,
3154
+
3155
+ 1
3156
+
3157
+ +
3158
+
3159
+ (
3160
+
3161
+ R R
3162
+ s
3163
+ p
3164
+
3165
+ )
3166
+
3167
+ and
3168
+
3169
+ (
3170
+
3171
+ 20
3172
+
3173
+ )
3174
+
3175
+ ´
3176
+
3177
+ (
3178
+
3179
+ a R
3180
+ s
3181
+
3182
+ )
3183
+
3184
+ 2
3185
+
3186
+ - +
3187
+ [
3188
+ 1
3189
+
3190
+ (
3191
+
3192
+ R R
3193
+ s
3194
+ p
3195
+
3196
+ 2
3197
+ )]
3198
+
3199
+ [
3200
+
3201
+ 1
3202
+
3203
+ +
3204
+
3205
+ (
3206
+
3207
+ R R
3208
+ s
3209
+ p
3210
+
3211
+ 2
3212
+ )]
3213
+
3214
+ -
3215
+
3216
+ 2
3217
+
3218
+ b
3219
+
3220
+
3221
+ d
3222
+ tr
3223
+ a R
3224
+ (
3225
+ s
3226
+
3227
+ )
3228
+
3229
+
3230
+
3231
+ = -
3232
+
3233
+ P
3234
+
3235
+ 24
3236
+ p
3237
+
3238
+ (
3239
+ a R
3240
+ s
3241
+
3242
+ a R
3243
+ )
3244
+ s
3245
+ 2
3246
+ -
3247
+ )
3248
+
3249
+ (
3250
+
3251
+ 2
3252
+
3253
+ b
3254
+
3255
+ [
3256
+
3257
+ 1
3258
+
3259
+ +
3260
+
3261
+ (
3262
+
3263
+ R R
3264
+ s
3265
+ p
3266
+
3267
+ 2
3268
+ )]
3269
+
3270
+ -
3271
+
3272
+ 2
3273
+
3274
+ b
3275
+
3276
+ (
3277
+
3278
+ a R
3279
+ s
3280
+
3281
+ )
3282
+
3283
+ 2
3284
+
3285
+ - +
3286
+ [
3287
+ 1
3288
+
3289
+ (
3290
+
3291
+ R R
3292
+ s
3293
+ p
3294
+
3295
+ 2
3296
+ )]
3297
+
3298
+
3299
+ d
3300
+ tr
3301
+
3302
+ b
3303
+
3304
+ = -
3305
+
3306
+ P
3307
+
3308
+ 24
3309
+ p
3310
+
3311
+ b
3312
+ 2
3313
+ )
3314
+
3315
+ (
3316
+
3317
+ a R
3318
+ s
3319
+
3320
+ -
3321
+
3322
+ 2
3323
+
3324
+ b
3325
+
3326
+ (
3327
+
3328
+ a R
3329
+ s
3330
+
3331
+ )
3332
+
3333
+ 2
3334
+
3335
+ - +
3336
+ [
3337
+ 1
3338
+
3339
+ (
3340
+
3341
+ R R
3342
+ s
3343
+ p
3344
+
3345
+ 2
3346
+ )]
3347
+
3348
+ [
3349
+
3350
+ 1
3351
+
3352
+ +
3353
+
3354
+ (
3355
+
3356
+ R R
3357
+ s
3358
+ p
3359
+
3360
+ 2
3361
+ )]
3362
+
3363
+ -
3364
+
3365
+ 2
3366
+
3367
+ b
3368
+
3369
+ ,
3370
+
3371
+ (
3372
+
3373
+ 21
3374
+
3375
+ )
3376
+
3377
+
3378
+ d
3379
+ in
3380
+
3381
+ P
3382
+
3383
+ =
3384
+
3385
+ d
3386
+ in
3387
+ P
3388
+
3389
+ ,
3390
+
3391
+
3392
+ d
3393
+ in
3394
+ R R
3395
+ p
3396
+
3397
+
3398
+
3399
+ (
3400
+
3401
+ =
3402
+
3403
+ P
3404
+
3405
+ 12
3406
+ p
3407
+
3408
+ )
3409
+ s
3410
+ ⎛
3411
+ ⎜
3412
+ ⎜
3413
+ ⎝
3414
+
3415
+ 1
3416
+
3417
+ +
3418
+
3419
+ (
3420
+
3421
+ R R
3422
+ p
3423
+
3424
+ s
3425
+
3426
+ )
3427
+
3428
+ (
3429
+
3430
+ a R
3431
+
3432
+ s
3433
+
3434
+ 2
3435
+
3436
+ )
3437
+
3438
+ - +
3439
+ [
3440
+ 1
3441
+
3442
+ (
3443
+
3444
+ R R
3445
+ p
3446
+
3447
+ s
3448
+
3449
+ 2
3450
+ )]
3451
+
3452
+ [
3453
+
3454
+ 1
3455
+
3456
+ +
3457
+
3458
+ (
3459
+
3460
+ R R
3461
+ p
3462
+
3463
+ s
3464
+
3465
+ 2
3466
+ )]
3467
+
3468
+ +
3469
+
3470
+ 1
3471
+
3472
+ -
3473
+
3474
+ (
3475
+
3476
+ R R
3477
+ p
3478
+
3479
+ s
3480
+
3481
+ )
3482
+
3483
+ (
3484
+
3485
+ a R
3486
+
3487
+ s
3488
+
3489
+ 2
3490
+
3491
+ )
3492
+
3493
+ - -
3494
+ [
3495
+ 1
3496
+
3497
+ (
3498
+
3499
+ R R
3500
+ p
3501
+
3502
+ s
3503
+
3504
+ 2
3505
+ )]
3506
+
3507
+ [
3508
+
3509
+ 1
3510
+
3511
+ -
3512
+
3513
+ (
3514
+
3515
+ R R
3516
+ p
3517
+
3518
+ s
3519
+
3520
+ 2
3521
+ )]
3522
+
3523
+ -
3524
+
3525
+ 2
3526
+
3527
+ b
3528
+
3529
+
3530
+ d
3531
+ in
3532
+ a R
3533
+ (
3534
+ s
3535
+
3536
+ )
3537
+
3538
+
3539
+
3540
+ = -
3541
+
3542
+ P
3543
+
3544
+ 12
3545
+ p
3546
+
3547
+ (
3548
+ a R
3549
+ s
3550
+
3551
+ a R
3552
+ )
3553
+ s
3554
+ 2
3555
+ -
3556
+ )
3557
+
3558
+ (
3559
+
3560
+ 2
3561
+
3562
+ b
3563
+
3564
+ ⎛
3565
+ ⎜
3566
+ ⎜
3567
+ ⎝
3568
+
3569
+ -
3570
+
3571
+ [
3572
+
3573
+ 1
3574
+
3575
+ -
3576
+
3577
+ (
3578
+
3579
+ R R
3580
+ s
3581
+ p
3582
+
3583
+ 2
3584
+ )]
3585
+
3586
+ -
3587
+
3588
+ 2
3589
+
3590
+ b
3591
+
3592
+ (
3593
+
3594
+ a R
3595
+ s
3596
+
3597
+ )
3598
+
3599
+ 2
3600
+
3601
+ - -
3602
+ [
3603
+ 1
3604
+
3605
+ (
3606
+
3607
+ R R
3608
+ s
3609
+ p
3610
+
3611
+ 2
3612
+ )]
3613
+
3614
+ [
3615
+
3616
+ 1
3617
+
3618
+ +
3619
+
3620
+ (
3621
+
3622
+ R R
3623
+ s
3624
+ p
3625
+
3626
+ 2
3627
+ )]
3628
+
3629
+ -
3630
+
3631
+ 2
3632
+
3633
+ b
3634
+
3635
+ (
3636
+
3637
+ a R
3638
+ s
3639
+
3640
+ - +
3641
+ [
3642
+ 1
3643
+
3644
+ (
3645
+
3646
+ R R
3647
+ s
3648
+ p
3649
+
3650
+ 2
3651
+ )]
3652
+
3653
+ 2
3654
+
3655
+ )
3656
+ ⎞
3657
+ ⎟
3658
+ ⎟
3659
+ ⎠
3660
+
3661
+ ,
3662
+
3663
+ (
3664
+
3665
+ 25
3666
+
3667
+ )
3668
+
3669
+ (
3670
+
3671
+ 26
3672
+
3673
+ )
3674
+
3675
+ ,
3676
+
3677
+ (
3678
+
3679
+ 27
3680
+
3681
+ )
3682
+
3683
+ ,
3684
+
3685
+ (
3686
+
3687
+ 28
3688
+
3689
+ )
3690
+
3691
+ (
3692
+
3693
+ 29
3694
+
3695
+ )
3696
+
3697
+ (
3698
+
3699
+ 30
3700
+
3701
+ )
3702
+
3703
+ ,
3704
+
3705
+ 2
3706
+
3707
+ -
3708
+
3709
+ b
3710
+ ⎞
3711
+ ⎟
3712
+ ⎟
3713
+ ⎠
3714
+
3715
+ (
3716
+
3717
+ 31
3718
+
3719
+ )
3720
+
3721
+ (
3722
+
3723
+ 32
3724
+
3725
+ )
3726
+
3727
+ Note that
3728
+
3729
+ the derived parameters i, dtr, din, and D are
3730
+ determined independently of the stellar parameters; therefore,
3731
+ their partial derivatives with respect to the stellar parameters are
3732
+ all identically zero.
3733
+
3734
+ As the transit depth D is determined from the model light
3735
+ curve generated by the geometric transit signal generator, the
3736
+ ¶ ¶(
3737
+ ¶(
3738
+ elements D R Rp
3739
+ the
3740
+ s
3741
+ ¶ are determined numerically. The other non-
3742
+ Jacobian y q
3743
+
3744
+ zero elements of the Jacobians y q
3745
+ are
3746
+ calculated according to the following equations:
3747
+
3748
+ ), and ∂D/∂b of
3749
+
3750
+ ¶ , and y a
3751
+
3752
+
3753
+ ), D
3754
+
3755
+ a Rs
3756
+
3757
+
3758
+
3759
+
3760
+
3761
+
3762
+
3763
+
3764
+
3765
+ R
3766
+
3767
+ p
3768
+ R R
3769
+ s
3770
+ p
3771
+
3772
+
3773
+
3774
+ (
3775
+
3776
+ =
3777
+
3778
+ )
3779
+
3780
+ R
3781
+
3782
+ p
3783
+ R R
3784
+ s
3785
+ p
3786
+
3787
+ (
3788
+
3789
+ )
3790
+
3791
+ ,
3792
+
3793
+ (
3794
+
3795
+ 22
3796
+
3797
+ )
3798
+
3799
+ 27
3800
+
3801
+ Publications of the Astronomical Society of the Pacific, 131:024506 (28pp), 2019 February
3802
+
3803
+ Li et al.
3804
+
3805
+
3806
+ d
3807
+ in
3808
+
3809
+ b
3810
+
3811
+ = -
3812
+
3813
+ P
3814
+
3815
+ 12
3816
+ p
3817
+
3818
+ b
3819
+ 2
3820
+ )
3821
+
3822
+ (
3823
+
3824
+ a R
3825
+ s
3826
+
3827
+ -
3828
+
3829
+ 2
3830
+
3831
+ b
3832
+
3833
+ ⎛
3834
+ ⎜
3835
+ ⎜
3836
+ ⎝
3837
+
3838
+ [
3839
+
3840
+ 1
3841
+
3842
+ (
3843
+
3844
+ a R
3845
+ s
3846
+
3847
+ )
3848
+
3849
+ 2
3850
+
3851
+ - +
3852
+ [
3853
+ 1
3854
+
3855
+ (
3856
+
3857
+ R R
3858
+ s
3859
+ p
3860
+
3861
+ 2
3862
+ )]
3863
+
3864
+ ORCID iDs
3865
+
3866
+ Joseph D. Twicken
3867
+
3868
+ https://orcid.org/0000-0002-6778-7552
3869
+
3870
+ References
3871
+
3872
+ Akeson, R. L., Chen, X., Ciardi, D., et al. 2013, PASP, 125, 989
3873
+ Brown, T. M., Latham, D. W., Everett, M. E., & Esquerdo, G. A. 2011, AJ,
3874
+
3875
+ 142, 112
3876
+
3877
+ Christiansen, J. L., Clarke, B. D., Burke, C. J., et al. 2013, ApJS, 207, 35
3878
+ Christiansen, J. L., Clarke, B. D., Burke, C. J., et al. 2015, ApJ, 810, 95
3879
+ Claret, A., & Bloemen, S. 2011, A&A, 529, AA75
3880
+ Crossfield, I. J. M., Ciardi, D. R., Petigura, E. A., et al. 2016, ApJS, 226, 7
3881
+ Crossfield, I. J. M., Guerrero, N., David, T., et al. 2018, ApJS, 239, 5
3882
+ Dressing, C. D., & Charboneau, D. 2015, ApJ, 807, 45
3883
+ Foreman-Mackey, D., Montet, B. T., Hogg, D. W., et al. 2015, ApJ, 806, 215
3884
+ Garcia, D. 2010, CSDA, 54, 1167
3885
+ Holland, P. W., & Welsch, R. E. 1977, Commun. Stat.—Theory Methods,
3886
+
3887
+ A6, 813
3888
+
3889
+ Jenkins, J. M. 2002, ApJ, 575, 493
3890
+ Jenkins, J. M., Caldwell, D. A., Chandrasekaran, H., et al. 2010a, ApJL,
3891
+
3892
+ 713, L87
3893
+
3894
+ Jenkins, J. M., Chandrasekaran, H., McCauliff, S. D., et al. 2010b, Proc. SPIE,
3895
+
3896
+ 7740, 77400D
3897
+
3898
+ Jenkins, J. M., Tenenbaum, P., Seader, S., et al. 2017, in Kepler Data Processing
3899
+ Handbook: Transit Planet Search (KSCI-19801-002), ed. J. M. Jenkins,
3900
+ https://archive.stsci.edu/kepler/manuals/KSCI-19081-002-KDPH.pdf
3901
+
3902
+ Levenberg, K. 1944, Q. Appl. Math., 2, 164
3903
+ Mandel, K., & Agol, E. 2002, ApJL, 580, L171
3904
+ Marquardt, D. W. 1963, J. Soc. Ind. Appl. Math., 11, 431
3905
+ Mathur, S., Huber, D., Batalha, N. M., et al. 2017, ApJS, 229, 30
3906
+ Petigura, E. A., Crossfield, I. J. M., Isaacson, H., et al. 2018, AJ, 155, 21
3907
+ Rizzuto, A. C., Mann, A. W., Vanderburg, A., et al. 2017, AJ, 154, 224
3908
+ Rowe, J. F., Bryson, S. T., Marcy, G. W., et al. 2014, ApJ, 784, 45
3909
+ Seader, S., Jenkins, J. M., Tenenbaum, P., et al. 2015, ApJS, 217, 18
3910
+ Smith, J. C., Stumpe, M. C., Van Cleve, J. E., et al. 2012, PASP, 124, 1000
3911
+ Stumpe, M. C., Smith, J. C., Van Cleve, J. E., et al. 2012, PASP, 124, 985
3912
+ Tenenbaum, P., Bryson, S. T., Chandrasekaran, H., et al. 2010, Proc. SPIE,
3913
+
3914
+ 7740, 77400J
3915
+
3916
+ Tenenbaum, P., Christiansen, J. L., Jenkins, J. M., et al. 2012, ApJS, 199, 24
3917
+ Tenenbaum, P., Jenkins, J. M., Seader, S., et al. 2013, ApJS, 206, 5
3918
+ Tenenbaum, P., Jenkins, J. M., Seader, S., et al. 2014, ApJS, 211, 6
3919
+ Thompson, S. E., Coughlin, J. L., Hoffman, K., et al. 2018, ApJS, 235, 38
3920
+ Twicken, J. D., Catanzarite, J. H., Clarke, B. D., et al. 2018, PASP, 130,
3921
+
3922
+ 064502
3923
+
3924
+ Twicken, J. D., Jenkins, J. M., Seader, S. E., et al. 2016, AJ, 152, 158
3925
+ Vanderburg, A., Latham, D. W., Buchhave, L. A., et al. 2016, ApJS, 222, 14
3926
+ Witteborn, F. C., Van Cleve, J., Borucki, W., Argabright, V., & Hascall, P.
3927
+
3928
+ 2011, Proc. SPIE, 8151, 815117
3929
+
3930
+ Wu, H., Twicken, J. D., Tenenbaum, P., et al. 2010, Proc. SPIE, 7740, 774019
3931
+ Yu, L., Crossfield, I. J. M., Schlieder, J. E., et al. 2018, AJ, 156, 22
3932
+
3933
+ -
3934
+
3935
+ (
3936
+
3937
+ a R
3938
+ s
3939
+
3940
+ )
3941
+
3942
+ 2
3943
+
3944
+ - -
3945
+ [
3946
+ 1
3947
+
3948
+ (
3949
+
3950
+ R R
3951
+ s
3952
+ p
3953
+
3954
+ 2
3955
+ )]
3956
+
3957
+ [
3958
+
3959
+ 1
3960
+
3961
+ -
3962
+
3963
+ (
3964
+
3965
+ R R
3966
+ s
3967
+ p
3968
+
3969
+ 2
3970
+ )]
3971
+
3972
+ -
3973
+
3974
+ 2
3975
+
3976
+ b
3977
+
3978
+
3979
+
3980
+ T
3981
+ eq
3982
+ P
3983
+
3984
+
3985
+
3986
+
3987
+
3988
+ f
3989
+ eff
3990
+ P
3991
+
3992
+
3993
+
3994
+ = -
3995
+
3996
+ = -
3997
+
3998
+ 1
3999
+ 3
4000
+
4001
+ 4
4002
+ 3
4003
+
4004
+ T
4005
+ eq
4006
+ P
4007
+
4008
+ ,
4009
+
4010
+ f
4011
+ eff
4012
+ P
4013
+
4014
+ ,
4015
+
4016
+
4017
+
4018
+ R
4019
+
4020
+ p
4021
+
4022
+ R
4023
+ s
4024
+
4025
+ =
4026
+
4027
+ R
4028
+
4029
+ p
4030
+ R
4031
+ s
4032
+
4033
+ ,
4034
+
4035
+
4036
+ a
4037
+ R
4038
+ s
4039
+
4040
+
4041
+ a
4042
+ g
4043
+
4044
+
4045
+
4046
+ T
4047
+ eq
4048
+
4049
+ R
4050
+ s
4051
+
4052
+ =
4053
+
4054
+ 2
4055
+ 3
4056
+
4057
+ a
4058
+ R
4059
+ s
4060
+
4061
+ ,
4062
+
4063
+ =
4064
+
4065
+ =
4066
+
4067
+ 1
4068
+ 3
4069
+
4070
+ 1
4071
+ 6
4072
+
4073
+ a
4074
+ g
4075
+
4076
+ ,
4077
+
4078
+ T
4079
+ eq
4080
+ R
4081
+ s
4082
+
4083
+ ,
4084
+
4085
+
4086
+
4087
+ T
4088
+ eq
4089
+ g
4090
+
4091
+ = -
4092
+
4093
+ 1
4094
+ 6
4095
+
4096
+ T
4097
+ eq
4098
+ g
4099
+
4100
+ ,
4101
+
4102
+
4103
+
4104
+ T
4105
+ eq
4106
+
4107
+ T
4108
+ eff
4109
+
4110
+ =
4111
+
4112
+ T
4113
+ eq
4114
+
4115
+ T
4116
+ eff
4117
+
4118
+ ,
4119
+
4120
+
4121
+
4122
+ f
4123
+ eff
4124
+
4125
+ R
4126
+ s
4127
+
4128
+ =
4129
+
4130
+ 2
4131
+ 3
4132
+
4133
+ f
4134
+ eff
4135
+ R
4136
+ s
4137
+
4138
+ ,
4139
+
4140
+
4141
+
4142
+ f
4143
+ eff
4144
+ g
4145
+
4146
+ = -
4147
+
4148
+ 2
4149
+ 3
4150
+
4151
+ f
4152
+ eff
4153
+ g
4154
+
4155
+ , and
4156
+
4157
+
4158
+
4159
+ f
4160
+ eff
4161
+
4162
+ T
4163
+ eff
4164
+
4165
+ =
4166
+
4167
+ 4
4168
+
4169
+ f
4170
+ eff
4171
+ T
4172
+ eff
4173
+
4174
+ .
4175
+
4176
+ (
4177
+
4178
+ R R
4179
+ s
4180
+ p
4181
+
4182
+ 2
4183
+ )]
4184
+
4185
+ -
4186
+
4187
+ 2
4188
+
4189
+ b
4190
+
4191
+ +
4192
+ ⎞
4193
+ ⎟
4194
+ ⎟
4195
+ ⎠
4196
+
4197
+ ,
4198
+
4199
+ (
4200
+
4201
+ 33
4202
+
4203
+ )
4204
+
4205
+ (
4206
+
4207
+ 34
4208
+
4209
+ )
4210
+
4211
+ (
4212
+
4213
+ 35
4214
+
4215
+ )
4216
+
4217
+ (
4218
+
4219
+ 36
4220
+
4221
+ )
4222
+
4223
+ (
4224
+
4225
+ 37
4226
+
4227
+ )
4228
+
4229
+ (
4230
+
4231
+ 38
4232
+
4233
+ )
4234
+
4235
+ (
4236
+
4237
+ 39
4238
+
4239
+ )
4240
+
4241
+ (
4242
+
4243
+ 40
4244
+
4245
+ )
4246
+
4247
+ (
4248
+
4249
+ 41
4250
+
4251
+ )
4252
+
4253
+ (
4254
+
4255
+ 42
4256
+
4257
+ )
4258
+
4259
+ (
4260
+
4261
+ 43
4262
+
4263
+ )
4264
+
4265
+ (
4266
+
4267
+ 44
4268
+
4269
+ )
4270
+
4271
+ 28
4272
+
4273
+
documents/txt/iatson.txt ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Martín Dévora-Pajares has worked to create a new multi-input multi-branch neural network (NN) to classify exoplanet candidate
2
+ signals from the Kepler and TESS official NASA missions. This NN receives the complete folded and binned light curve,
3
+ three versions of the binned light curve (focused on the period, its harmonic and its sub-harmonic), the focused
4
+ folded and binned curve of centroids/motion data of the spacecraft and the folded and binned curve of the optical ghost
5
+ computed data between the core aperture and its halo aperture. The NN also receives a few numerical values indicating
6
+ the quality of the transit source (number of complete transits, percentage of complete transits), the transit depth and
7
+ the planetary computed equilibrium temperature. Each of the folded-curve inputs is passed through an independent
8
+ convolutional branch and then, all the outputs are processed together in dense layers, returning a final only value
9
+ between 0 and 1 to indicate the quality of the signal (where 1 is a good one).
10
+
11
+ The training data is prepared by computing all the input data before the training stage so the neural network can ingest it fastly
12
+ from the disk. The data is labeled as `candidate`, `false positive`, `false alarm`, `eclipsing binary`,
13
+ `background eclipsing binary`, `confirmed planet`, `confirmed planet with radial velocities`, `validated planet`,
14
+ `tce` (threshold crossing event). Signals marked with `candidate` are excluded from the training set, as well as the
15
+ `validated planet` ones as they are statistically confirmed with similar tools and should not be used as input for a new
16
+ numerical or machine learning method. All the curves are detrended with a duration of 4 times the transit duration to ensure
17
+ that the transit alike signals are not removed.
18
+
19
+ To reach a high performant model we always need to register the relevant metrics that we want to focus on and select the
20
+ most relevant one to select the so called `best model`. In our case, we registered the accuracy, precision, recall,
21
+ recall at a precision of 0.99, where we used the last one to select the `best model`. We did it like this, as it was done
22
+ in Exominer (Valizadegan et al., 2022), because we wanted to ensure the highest recovery rate for a precision near to 1.
23
+ That is, we want a model whose positive predictions are almost never wrong, trying to leave as less real positives as we can
24
+ outside the true predictions. To reach a high recall at precision of 0.99, the root model structure is not enough to
25
+ achieve great performances. In order to get better results one need to use several regularization techniques, as the model
26
+ tends to overfit quickly to the training data, failing to generalize to unknown examples. There are many of these techniques
27
+ out there: dropout, noise addition, L1 and L2 regularizations, learning rate re-scaling, etc.
28
+
29
+ A second optional approach that is currently being evaluated is to create one model per type of false positive (binaries, optical ghosts, centroids, etc)
30
+ and later create an ensemble incorporating all those models. That way the final neural network would be very explainable, which is one of
31
+ the keystones in the field right now. The option of replacing the convolutional layers by transformers is also being evaluated, as
32
+ they have overpowered the first ones in almost every field where they have been used. Transformers have the capability of relate
33
+ all the data in the provided window instead of creating features of nearby neurons, as the convolutional layers do. They require
34
+ more computational resources, though, and are harder to fine-tune.
documents/txt/matrix.txt ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MATRIX ToolKit
2
+ ToolKit for Multi-phAse Transits Recovery from Injected eXoplanets
3
+
4
+ MATRIX is an injection-recovery software prepared to create grids of scenarios with a set of periods, radii and epochs of synthetic transiting exoplanet signals in a provided light curve. The typical injection-recovery execution has consisted in the usage of 2 dimensional scenarios, where only one epoch (random or hardcoded) was used for each period and radius. Even though this is a fair approach when the computational power is very limited, this reduces the accuracy of the experiment as the recoverability might vary depending on the transit epoch and hence, the places where the transits would appear in the curve. E.g: A transit would be clearly recovered falling in a quite region of the curve but not easy to recover when it happened close to a flare or a curve gap. A multi-phase analysis can now be done with MATRIX easily by only setting a few parameters in a configuration file and running one line of code.
5
+
6
+ Just execute the command below this text. Take into accont that the user-properties.yaml file needs to include several mandatory options. Please refer to the example files under the example directory.
7
+
8
+ python3.8 -m tkmatrix --properties user-properties.yaml
9
+
10
+ For the complete set of properties available for use please look at the root properties.yaml
11
+
12
+ By-products
13
+ a_tls_report.csv: A file containing a csv formatted output given the orbital period, the radius and the epoch besides the outputs with found status, SNR and SDE of the results.
14
+ a_tls_report.png: A file with an automatically generated plot from the csv report. You are free to build your own plot from the report if you feel like the one provided by MATRIX is not good enough for your purposes.
15
+ Injected curves (csv files): In case you want to study the injected curves generated for the recovery, you can set a flag to the tool so that it keeps the files after it finishes. If you don't provide that flag, the files will be removed at the end of the execution.
16
+
17
+ Citation
18
+ We are planning to write a scientific paper based on the usage of MATRIX. In the meantime, we encourage the users to cite the Software DOI in their research:
19
+
20
+ @MISC{2022zndo...6570831D,
21
+ author = {{D{\'e}vora-Pajares}, Mart{\'\i}n and {Pozuelos}, Francisco J.},
22
+ title = "{MATRIX: Multi-phAse Transits Recovery from Injected eXoplanets}",
23
+ keywords = {exoplanets, transits, injection \& recovery, python},
24
+ howpublished = {Zenodo},
25
+ year = 2022,
26
+ month = may,
27
+ eid = {10.5281/zenodo.6570831},
28
+ doi = {10.5281/zenodo.6570831},
29
+ version = {0.3.17},
30
+ publisher = {Zenodo},
31
+ adsurl = {https://ui.adsabs.harvard.edu/abs/2022zndo...6570831D},
32
+ adsnote = {Provided by the SAO/NASA Astrophysics Data System}
33
+ }
34
+
35
+ Documentation
36
+ For more information please visit https://tkmatrix.readthedocs.io/.
documents/txt/publications.txt ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ This is the list of publications where Martín Dévora-Pajares has collaborated
2
+ [1]Dévora-Pajares, M., Pozuelos, F. J., Thuillier, A., García, L., and Cerdeño Mota, L., “SHERLOCK: Explore Kepler, K2, and TESS data”, <i>Astrophysics Source Code Library</i>, 2023. ascl:2306.043.
3
+ [2]Dransfield, G., “A 1.55 R<SUB>⨁</SUB> habitable-zone planet hosted by TOI-715, an M4 star near the ecliptic South Pole”, <i>Monthly Notices of the Royal Astronomical Society</i>, 2023. doi:10.1093/mnras/stad1439.
4
+ [3]Pozuelos, F. J., “A super-Earth and a mini-Neptune near the 2:1 MMR straddling the radius valley around the nearby mid-M dwarf TOI-2096”, <i>Astronomy and Astrophysics</i>, vol. 672, 2023. doi:10.1051/0004-6361/202245440.
5
+ [4]Dransfield, G., “HD 28109 hosts a trio of transiting Neptunian planets including a near-resonant pair, confirmed by ASTEP from Antarctica”, <i>Monthly Notices of the Royal Astronomical Society</i>, vol. 515, no. 1, pp. 1328–1345, 2022. doi:10.1093/mnras/stac1383.
6
+ [5]Pozuelos, F. J., Dévora-Pajares, M., Thuillier, A., Van Grootel, V., and Suarez Yanes, J. C., “SHERLOCK: A python pipeline to explore space-based observations in the search for planets”, 2022. doi:10.5194/epsc2022-970.
7
+ [6]Thuillier, A., Van Grootel, V., Pozuelos, F., Devora-Pajares, M., Siess, L., and Charpinet, S., “A search for transiting planets around hot subdwarfs - Results from TESS Cycle I”, 2022. doi:10.5194/epsc2022-514.
8
+ [7]Thuillier, A., Van Grootel, V., Dévora-Pajares, M., Pozuelos, F. J., Charpinet, S., and Siess, L., “A search for transiting planets around hot subdwarfs. II. Supplementary methods and results from TESS Cycle 1”, <i>Astronomy and Astrophysics</i>, vol. 664, 2022. doi:10.1051/0004-6361/202243554.
9
+ [8]Thuillier, A., Pozuelos, F. J., Dévora-Pajares, M., Charpinet, S., and Siess, L., “A transit survey to search for planets around hot subdwarfs %t À la recherche de transits planétaires autour d'étoiles sous-naines chaudes”, <i>Bulletin de la Societe Royale des Sciences de Liege</i>, vol. 91, no. 2, 2022.
10
+ [9]Pozuelos, F. J., Van Grootel, V., Thuillier, A., and Dévora-Pajares, M. M., “The search for transiting planets around hot subdwarf”, <i>Bulletin de la Societe Royale des Sciences de Liege</i>, vol. 91, no. 2, 2022.
11
+ [10]Dévora-Pajares, M. and Pozuelos, F. J., “MATRIX: Multi-phAse Transits Recovery from Injected eXoplanets”, in <i>Zenodo Software package</i>, 2022, vol. 65. doi:10.5281/zenodo.6570831.
12
+ [11]Schanche, N., “TOI-2257 b: A highly eccentric long-period sub-Neptune transiting a nearby M dwarf”, <i>Astronomy and Astrophysics</i>, vol. 657, 2022. doi:10.1051/0004-6361/202142280.
13
+ [12]Dévora-Pajares, M., Francisco J. Pozuelos, Juan C. Suárez, and Garanat Team, “The SHERLOCK PIPEline: Searching for Hints of Exoplanets fRom Lightcurves Of spaCe-based seeKers”, in <i>Plato Mission Conference 2021. Presentations and posters of the online PLATO Mission Conference 2021</i>, 2021. doi:10.5281/zenodo.5557904.
14
+ [13]Wells, R. D., “A large sub-Neptune transiting the thick-disk M4 V TOI-2406”, <i>Astronomy and Astrophysics</i>, vol. 653, 2021. doi:10.1051/0004-6361/202141277.
15
+ [14]Thuillier, A., Van Grootel, V., Pozuelos, F., Dévora-Pajares, M., and Charpinet, S., “A transit survey to search for planets around hot subdwarfs”, in <i>Posters from the TESS Science Conference II (TSC2)</i>, 2021. doi:10.5281/zenodo.5118657.
16
+ [15]Van Grootel, V., “A search for transiting planets around hot subdwarfs. I. Methods and performance tests on light curves from Kepler, K2, TESS, and CHEOPS”, <i>Astronomy and Astrophysics</i>, vol. 650, 2021. doi:10.1051/0004-6361/202140381.
17
+ [16]Van Grootel, V., “VizieR Online Data Catalog: Search for transiting planets around sdBs (Van Grootel+, 2021)”, <i>VizieR Online Data Catalog</i>, 2021.
documents/txt/sherlock.txt ADDED
@@ -0,0 +1,162 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ SHERLOCK is an end-to-end pipeline that allows the users to explore the data from space-based missions to search for planetary candidates. It can be used to recover alerted candidates by the automatic pipelines such as SPOC and the QLP, the so-called Kepler objects of interest (KOIs) and TESS objects of interest (TOIs), and to search for candidates that remain unnoticed due to detection thresholds, lack of data exploration or poor photometric quality. To this end, SHERLOCK has six different modules to (1) acquire and prepare the light curves from their repositories, (2) search for planetary candidates, (3) vet the interesting signals, (4) perform a statistical validation, (5) model the signals to refine their ephemerides, and (6) compute the observational windows from ground-based observatories to trigger a follow-up campaign. To execute all these modules, the user only needs to fill in an initial YAML file with some basic information such as the star ID (KIC-ID, EPIC-ID, TIC-ID), the cadence to be used, etc., and use sequentially a few lines of code to pass from one step to the next. Alternatively, the user may provide with the light curve in a csv file, where the time, the normalized flux, and the flux error need to be given in columns comma-separated format.
2
+
3
+ Citation
4
+ We are currently working on a specific paper for SHERLOCK. In the meantime, the best way to cite SHERLOCK is by referencing the first paper where it was used Pozuelos et al. (2020):
5
+
6
+ @ARTICLE{2020A&A...641A..23P,
7
+ author = {{Pozuelos}, Francisco J. and {Su{\'a}rez}, Juan C. and {de El{\'\i}a}, Gonzalo C. and {Berdi{\~n}as}, Zaira M. and {Bonfanti}, Andrea and {Dugaro}, Agust{\'\i}n and {Gillon}, Micha{\"e}l and {Jehin}, Emmanu{\"e}l and {G{\"u}nther}, Maximilian N. and {Van Grootel}, Val{\'e}rie and {Garcia}, Lionel J. and {Thuillier}, Antoine and {Delrez}, Laetitia and {Rod{\'o}n}, Jose R.},
8
+ title = "{GJ 273: on the formation, dynamical evolution, and habitability of a planetary system hosted by an M dwarf at 3.75 parsec}",
9
+ journal = {\aap},
10
+ keywords = {planets and satellites: dynamical evolution and stability, planets and satellites: formation, Astrophysics - Earth and Planetary Astrophysics, Astrophysics - Solar and Stellar Astrophysics},
11
+ year = 2020,
12
+ month = sep,
13
+ volume = {641},
14
+ eid = {A23},
15
+ pages = {A23},
16
+ doi = {10.1051/0004-6361/202038047},
17
+ archivePrefix = {arXiv},
18
+ eprint = {2006.09403},
19
+ primaryClass = {astro-ph.EP},
20
+ adsurl = {https://ui.adsabs.harvard.edu/abs/2020A&A...641A..23P},
21
+ adsnote = {Provided by the SAO/NASA Astrophysics Data System}
22
+ }
23
+
24
+ Also, you may be interested in having a look at recent papers that used SHERLOCK:
25
+ Pozuelos et al. (2023)
26
+ Delrez et al. (2022)
27
+ Dransfield et al. (2022)
28
+ Luque et al. (2022)
29
+ Schanche et al. (2022)
30
+ Wells et al. (2021)
31
+ Benni et al. (2021)
32
+ Van Grootel et al. (2021)
33
+ Demory et al. (2020)
34
+
35
+ Main Developers
36
+ Active: F.J. Pozuelos, M. Dévora
37
+
38
+ Additional contributors
39
+ A. Thuillier & L. García & Luis Cerdeño Mota
40
+
41
+ Documentation
42
+ Please visit https://sherlock-ph.readthedocs.io to get a complete set of explanations and tutorials to get started with SHERLOCK.
43
+
44
+ Launch
45
+ You can run SHERLOCK PIPEline as a standalone package by using:
46
+
47
+ python3 -m sherlockpipe --properties my_properties.yaml
48
+
49
+ You only need to provide a YAML file with any of the properties contained in the internal properties.yaml provided by the pipeline. The most important keys to be defined in your YAML file are those under the GLOBAL OBJECTS RUN SETUP and SECTOR OBJECTS RUN SETUP sections because they contain the object ids or files to be analysed in the execution. You'd need to fill at least one of those keys for the pipeline to do anything. If you still have any doubts please refer to the examples/properties directory
50
+
51
+ Additionally, you could only want to inspect the preparation stage of SHERLOCK and therefore, you can execute it without running the analyse phase so you can watch the light curve, the periodogram and the initial report to take better decisions to tune the execution parameters. Just launch SHERLOCK with:
52
+
53
+ python3 -m sherlockpipe --properties my_properties.yaml --explore
54
+
55
+ and it will end as soon as it has processed the preparation stages for each object.
56
+
57
+ Updates
58
+ SHERLOCK uses third party data to know TOIs, KOIs, EPICs and to handle FFIs and the vetting process. This data gets frequently updated from the active missions and therefore SHERLOCK will perform better if the metadata gets refreshed. You can simply run:
59
+
60
+ python3 -m sherlockpipe.update
61
+
62
+ and SHERLOCK will download the dependencies. It will store a timestamp to remember the last time it was refreshed to prevent several unneeded calls. However, if you find that there are more updates and you need them now, you can call:
63
+
64
+ python3 -m sherlockpipe.update --force
65
+
66
+ and SHERLOCK will ignore the timestamps and perform the update process. In addition, you could be interested in wiping all the metadata and build it again. That's why you could execute:
67
+
68
+ python3 -m sherlockpipe.update --clean
69
+
70
+ This last command implies a force statement and the last executed time will be ignored too.
71
+
72
+ You can additionally let SHERLOCK refresh the OIs list before running your current execution by adding to the YAML file the next line:
73
+
74
+ UPDATE_OIS=True
75
+
76
+ Vetting
77
+ SHERLOCK PIPEline comes with a submodule to examine the most promising transit candidates found by any of its executions. This is done via WATSON, capable of vetting TESS and Kepler targets. You should be able to execute the vetting by calling:
78
+
79
+ python3 -m sherlockpipe.vet --properties my_properties.yaml
80
+
81
+ Through that command you will run the vetting process for the given parameters within your provided YAML file. You could watch the generated results under $your_sherlock_object_results_dir/vetting directory. Please go to examples/vetting/ to learn how to inject the proper properties for the vetting process.
82
+
83
+ There is an additional simplified option which can be used to run the vetting. In case you are sure there is a candidate from the Sherlock results which matches your desired parameters, you can run
84
+
85
+ python3 -m sherlockpipe.vet --candidate ${theCandidateNumber}
86
+
87
+ from the sherlock results directory. This execution will automatically read the transit parameters from the Sherlock generated files.
88
+
89
+ Fitting
90
+ SHERLOCK PIPEline comes with another submodule to fit the most promising transit candidates found by any of its executions. This fit is done via ALLESFITTER code. By calling:
91
+
92
+ python3 -m sherlockpipe.fit --properties my_properties.yaml
93
+
94
+ you will run the fitting process for the given parameters within your provided YAML file. You could watch the generated results under $your_sherlock_object_results_dir/fit directory. Please go to examples/fitting/ to learn how to inject the proper properties for the fitting process.
95
+
96
+ There is an additional simplified option which can be used to run the fit. In case you are sure there is a candidate from the Sherlock results which matches your desired parameters, you can run
97
+
98
+ python3 -m sherlockpipe.fit --candidate ${theCandidateNumber}
99
+
100
+ from the sherlock results directory. This execution will automatically read the transit and star parameters from the Sherlock generated files.
101
+
102
+ Validation
103
+ SHERLOCK PIPEline implements a module to execute a statistical validation of a candidate by the usage of TRICERATOPS. By calling:
104
+
105
+ python3 -m sherlockpipe.validate --candidate ${theCandidateNumber}
106
+
107
+ you will run the validation for one of the Sherlock candidates.
108
+
109
+ Stability
110
+ SHERLOCK PIPEline also implements a module to execute a system stability computation by the usage of Rebound and SPOCK. By calling:
111
+
112
+ python3 -m sherlockpipe.stability --bodies 1,2,4
113
+
114
+ where the --bodies parameter is the set of the SHERLOCK accepted signals as CSV to be used in the scenarios simulation. You can also provide a stability properties file) to run a custom stability simulation:
115
+
116
+ python3 -m sherlockpipe.stability --properties stability.yaml
117
+
118
+ and you can even combine SHERLOCK accepted signals with some additional bodies provided by the properties file:
119
+
120
+ python3 -m sherlockpipe.stability --bodies 1,2,4 --properties stability.yaml
121
+
122
+ The results will be stored into a stability directory containing the execution log and a stability.csv containing one line per simulated scenario, sorted by the best results score.
123
+
124
+ Observation plan
125
+ SHERLOCK PIPEline also adds now a tool to plan your observations from ground-based observatories by using astropy and astroplan. By calling:
126
+
127
+ python3 -m sherlockpipe.plan --candidate ${theCandidateNumber} --observatories observatories.csv
128
+
129
+ on the resulting sherlockpipe.fit directory, where the precise candidate ephemeris are placed. The observatories.csv file should contain the list of available observatories for your candidate follow-up. As an example, you can look at this file.
130
+
131
+ SHERLOCK PIPEline Workflow
132
+ It is important to note that SHERLOCK PIPEline uses some csv files with TOIs, KOIs and EPIC IDs from the TESS, Kepler and K2 missions. Therefore your first execution of the pipeline might take longer because it will download the information.
133
+
134
+ Provisioning of light curve
135
+ The light curve for every input object needs to be obtained from its mission database. For this we use the high level API of Lightkurve, which enables the download of the desired light curves for TESS, Kepler and K2 missions. We also include Full Frame Images from the TESS mission by the usage of ELEANOR. We always use the PDCSAP signal from the ones provided by any of those two packages.
136
+
137
+ Pre-processing of light curve
138
+ In many cases we will find light curves which contain several systematics like noise, high dispersion beside the borders, high-amplitude periodicities caused by pulsators, fast rotators, etc. SHERLOCK PIPEline provides some methods to reduce these most important systematics.
139
+
140
+ Local noise reduction
141
+ For local noise, where very close measurements show high deviation from the local trend, we apply a Savitzky-Golay filter. This has proved a highly increment of the SNR of found transits. This feature can be disabled with a flag.
142
+
143
+ High RMS areas masking
144
+ Sometimes the spacecrafts have to perform reaction wheels momentum dumps by firing thrusters, sometimes there is high light scattering and sometimes the spacecraft can infer some jitter into the signal. For all of those systematics we found that in many cases the data from those regions should be discarded. Thus, SHERLOCK PIPEline includes a binned RMS computation where bins whose RMS value is higher than a configurable factor multiplied by the median get automatically masked. This feature can be disabled with a flag.
145
+
146
+ Input time ranges masking
147
+ If enabled, this feature automatically disables High RMS areas masking for the assigned object. The user can input an array of time ranges to be masked into the original signal.
148
+
149
+ Detrend of high-amplitude periodicities
150
+ Our most common foes with high periodicities are fast-rotators, which infer a high sinusoidal-like trend in the PDCSAP signal. This is why SHERLOCK PIPEline includes an automatic high-amplitude periodicities detection and detrending during its preparation stage. This feature can be disabled with a flag.
151
+
152
+ Input period detrend
153
+ If enabled, this feature automatically disables Detrend of high-amplitude periodicities for the assigned object. The user can input a period to be used for an initial detrend of the original signal.
154
+
155
+ Custom user code
156
+ You can even inject your own python code to perform:
157
+
158
+ A custom signal preparation task by implementing the CurvePreparer class that we provide. Then, inject your python file into the CUSTOM_PREPARER property and let SHERLOCK use your code.
159
+ A custom best signal selection algorithm by implementing the SignalSelector. class that we provide. Then, inject your python file into the CUSTOM_ALGORITHM property and let SHERLOCK use your code.
160
+ A custom search zone definition by implementing the SearchZone. class that we provide. Then, inject your python file into the CUSTOM_SEARCH_ZONE property and let SHERLOCK use your code.
161
+ Custom search modes: 'tls', 'bls', 'grazing', 'comet' or 'custom'. You can search for transits by using TLS, BLS, TLS for a grazing template, TLS for a comet template or even inject your custom transit template (this is currently included as an experimental feature).
162
+ For better understanding of usage please see the examples, which references custom implementations that you can inspect in our custom algorithms directory
documents/txt/toi2096.txt ADDED
The diff for this file is too large to render. See raw diff
 
documents/txt/toi2257.txt ADDED
@@ -0,0 +1,2603 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ A&A 657, A45 (2022)
2
+ https://doi.org/10.1051/0004-6361/202142280
3
+ © ESO 2022
4
+
5
+ TOI-2257 b: A highly eccentric long-period sub-Neptune
6
+ transiting a nearby M dwarf
7
+
8
+ N. Schanche1
9
+
10
+ , F. J. Pozuelos2,3, M. N. Günther4,5,(cid:63),(cid:63)(cid:63), R. D. Wells1, A. J. Burgasser6, P. Chinchilla2,7, L. Delrez2,3,
11
+
12
+ E. Ducrot2, L. J. Garcia2, Y. Gómez Maqueo Chew8, E. Jofré8,9,10, B. V. Rackham11,4,(cid:63)(cid:63)(cid:63), D. Sebastian12,
13
+ K. G. Stassun13, D. Stern14, M. Timmermans2, K. Barkaoui2,15, A. Belinski16, Z. Benkhaldoun15, W. Benz1,17,
14
+ A. Bieryla18, F. Bouchy19, A. Burdanov11, D. Charbonneau18, J. L. Christiansen20, K. A. Collins18,
15
+ B.-O. Demory1, M. Dévora-Pajares21, J. de Wit11, D. Dragomir22, G. Dransfield12, E. Furlan23, M. Ghachoui2,15,
16
+ M. Gillon2, C. Gnilka24, M. A. Gómez-Muñoz25, N. Guerrero26,4, M. Harris22, K. Heng1,27, C. E. Henze24,
17
+ K. Hesse4, S. B. Howell24, E. Jehin3, J. Jenkins24, E. L. N. Jensen28, M. Kunimoto4, D. W. Latham18, K. Lester24,
18
+ K. K. McLeod29, I. Mireles22, C. A. Murray30, P. Niraula11, P. P. Pedersen30, D. Queloz30, E. V. Quintana31,
19
+ G. Ricker4, A. Rudat4, L. Sabin25, B. Safonov16, U. Schroffenegger1, N. Scott24, S. Seager4,11,32, I. Strakhov16,
20
+ A. H. M. J. Triaud12, R. Vanderspek4, M. Vezie4, and J. Winn33
21
+
22
+ (Affiliations can be found after the references)
23
+
24
+ Received 22 September 2021 / Accepted 28 October 2021
25
+
26
+ ABSTRACT
27
+
28
+ Context. Thanks to the relative ease of finding and characterizing small planets around M-dwarf stars, these objects have become
29
+ cornerstones in the field of exoplanet studies. The current paucity of planets in long-period orbits around M dwarfs makes such objects
30
+ particularly compelling as they provide clues about the formation and evolution of these systems.
31
+ Aims. In this study we present the discovery of TOI-2257 b (TIC 198485881), a long-period (35 d) sub-Neptune orbiting an M3 star
32
+ at 57.8 pc. Its transit depth is about 0.4%, large enough to be detected with medium-size, ground-based telescopes. The long transit
33
+ 0.5), which would make it the most eccentric planet known to be
34
+ duration suggests the planet is in a highly eccentric orbit (e
35
+ transiting an M-dwarf star.
36
+ Methods. We combined TESS and ground-based data obtained with the 1.0-meter SAINT-EX, 0.60-meter TRAPPIST-North, and
37
+ 1.2-meter FLWO telescopes to find a planetary size of 2.2 R
38
+ and an orbital period of 35.19 days. In addition, we make use of archival
39
+ data, high-resolution imaging, and vetting packages to support our planetary interpretation.
40
+ Results. With its long period and high eccentricity, TOI-2257 b falls into a novel slice of parameter space. Despite the planet’s low
41
+ 0.015) and relative infrared brightness (Kmag = 10.7) make
42
+ equilibrium temperature (
43
+ it a suitable candidate for atmospheric exploration via transmission spectroscopy.
44
+
45
+ 256 K), its host star’s small size (R
46
+ ∗
47
+
48
+ = 0.311
49
+
50
+ ∼
51
+
52
+ ∼
53
+
54
+ ±
55
+
56
+ ⊕
57
+
58
+ Key words. planets and satellites: detection – stars: individual: TOI-2257 – stars: individual: TIC 198485881 –
59
+ techniques: photometric
60
+
61
+ 1. Introduction
62
+
63
+ Despite their cool temperatures, M dwarfs have become hot tar-
64
+ gets for exoplanet surveys. The large planet-to-star ratios of these
65
+ systems result in relatively deep transits and large radial veloc-
66
+ ity (RV) amplitudes. In addition, they emit more strongly in
67
+ the infrared. This leads to favorable conditions for atmospheric
68
+ characterization by the Hubble Space Telescope (HST) and the
69
+ James Webb Space Telescope (JWST). Several ground-based
70
+ surveys, such as SPECULOOS (Delrez et al. 2018; Sebastian
71
+ et al. 2021) and MEarth (Nutzman & Charbonneau 2008) via
72
+ the transit method and CARMENES (Reiners et al. 2018) and
73
+ the M-dwarf sample of HARPS (Bonfils et al. 2013) using RVs,
74
+ have been created with the specific goal of finding planets around
75
+ M dwarfs.
76
+
77
+ While space-based discoveries have been led by Kepler
78
+ (Borucki et al. 2010) and K2 (Howell et al. 2014), the Transiting
79
+ Exoplanet Survey Satellite (TESS; Ricker et al. 2015) is mak-
80
+ ing a growing contribution to the population of planet-hosting
81
+ M dwarfs, with 28 of 185 such planets discovered. Of those, only
82
+ 17 are transiting planets in long-period (>15 day) orbits1.
83
+
84
+ In this paper we present the discovery and statistical val-
85
+ idation of a long-period sub-Neptune orbiting an M 3 star.
86
+ Intriguingly, the planet’s orbit is highly elliptical, suggestive of
87
+ a possible perturbing outer gas giant (Huang et al. 2017; Van
88
+ Eylen et al. 2019). The planet was first identified in TESS data,
89
+ with further confirmation from several ground-based facilities,
90
+ including three telescopes in the SPECULOOS group (Sebastian
91
+ et al. 2021; Murray et al. 2020) that are designed for observations
92
+ of small planets around ultra-cool M dwarfs. The planet is fur-
93
+ ther validated with contributions from high-resolution imaging,
94
+ spectroscopy, and archival data.
95
+
96
+ (cid:63) Juan Carlos Torres Fellow.
97
+ (cid:63)(cid:63) ESA Research Fellow.
98
+ (cid:63)(cid:63)(cid:63) 51 Pegasi b Fellow.
99
+
100
+ 1 https://exoplanetarchive.ipac.caltech.edu/
101
+
102
+ Article published by EDP Sciences
103
+
104
+ A45, page 1 of 18
105
+
106
+ Astronomy&Astrophysics A&A 657, A45 (2022)
107
+
108
+ Fig. 1. Possible orbital solutions for the observed TESS transits. In-transit data points are shown in red. Based on the pattern of detections and
109
+ the observation spacing, possible orbital periods include 176, 88, 59, 44, and 35 days, shown by the red, orange, mustard, green, and blue ticks,
110
+ respectively.
111
+
112
+ Table 1. Ground-based photometric time-series observations of
113
+ TOI-2257.
114
+
115
+ Date (UT)
116
+
117
+ Filter
118
+
119
+ Facility
120
+
121
+ Exp. time (s)
122
+
123
+ Notes
124
+
125
+ 2021 April 20 I + z
126
+ 2021 April 20
127
+ 2021 April 20
128
+ 2021 May 25
129
+ 2021 May 25
130
+ 2021 May 25
131
+
132
+ i(cid:48)
133
+ z(cid:48)
134
+ I + z
135
+ r(cid:48)
136
+ z(cid:48)
137
+
138
+ SAINT-EX
139
+ KeplerCam
140
+ TRAPPIST-N
141
+ SAINT-EX
142
+ KeplerCam
143
+ KeplerCam
144
+
145
+ 20
146
+ 60
147
+ 60
148
+ 25
149
+ 120
150
+ 90
151
+
152
+ Partial (egress)
153
+ Partial (egress)
154
+ Partial (ingress)
155
+ Full
156
+ Partial (ingress)
157
+ Partial (ingress)
158
+
159
+ The structure of the present work is as follows. In Sect. 2
160
+ we introduce the contributing facilities and data sets used in the
161
+ validation of TOI-2257 b. Section 3 describes the work to charac-
162
+ terize the host star. The analysis of the transit light curves and the
163
+ characterization of the system are presented in Sect. 4. In Sect. 5
164
+ we discuss possible false-positive scenarios and argue that there
165
+ is sufficient evidence for the transits to be validated as planetary
166
+ in nature. The possibility of an additional planetary companion
167
+ in the system is explored in Sect. 6. Finally, implications for this
168
+ planet are discussed in Sect. 7.
169
+
170
+ 2. Observations
171
+
172
+ In this section we present all observations of TOI-2257 taken by
173
+ TESS and ground-based follow-up facilities. A summary of the
174
+ photometric follow-up observations is shown in Table 1.
175
+
176
+ 2.1. TESS photometry
177
+
178
+ TOI-2257 is a part of the TESS Candidate Target List (Stassun
179
+ et al. 2018b) and was observed with a 2-min cadence in TESS
180
+ Sectors 14 (2019 July 18–2019 August 16), 20 (2019 Decem-
181
+ ber 24–2020 January 21), 21 (2020 January 21–2020 February
182
+ 18), and 26 (2020 June 8–2020 July 4). The image data were
183
+ processed by the Science Processing Operations Center (SPOC)
184
+ pipeline (Jenkins et al. 2016) to produce photometric time series.
185
+ These time series were searched for transiting planet signatures
186
+ with a noise-compensating matched filter (Jenkins 2002; Jenkins
187
+ et al. 2010, 2020), which detected a pair of transits separated by
188
+ 175 days. The transit signature passed all the diagnostic tests per-
189
+ formed on the data (Twicken et al. 2018; Li et al. 2019), including
190
+ the difference image centroid test, which located the source of
191
+ 4.4 arcsec. No additional
192
+ the transit signature to within 1.5
193
+
194
+ ±
195
+
196
+ A45, page 2 of 18
197
+
198
+ transit signatures were found in the subsequent multiple planet
199
+ search. The TESS Science Office reviewed the data validation
200
+ reports and issued an alert for this planet candidate on 2020
201
+ August 10 (Guerrero et al. 2021).
202
+
203
+ Across the four sectors, only two transit events were detected.
204
+ The separation of the two observed transits and the gaps in cover-
205
+ age of the object led to a possible period of 175.9 days, but other
206
+ possibilities, including 88 d, 58.6 d, 44 d, and 35.2 d, could not
207
+ be ruled out (see Fig. 1). From the Mikulski Archive for Space
208
+ Telescopes (MAST) database, we obtained the Presearch Data
209
+ Conditioning Simple Aperture Photometry (PDC-SAP; Smith
210
+ et al. 2012; Stumpe et al. 2012, 2014), in which long-term trends
211
+ in the data are removed. In addition, we removed all data points
212
+ for which the quality flag was not 0. Figure 2 shows the apertures
213
+ used for the TESS data with Gaia Data Release 2 (DR2) sources
214
+ overlaid on the images.
215
+
216
+ 2.2. Follow-up photometry
217
+
218
+ Several ground-based observations of TOI-2257 were taken in
219
+ order to secure the period derived from TESS, constrain the
220
+ transit parameters, and characterize the star. A non-detection
221
+ on 2021 March 4 using the LCO McDonald 1-meter telescope
222
+ ruled out the 58-day alias. The 35-day alias was sampled next
223
+ on 2021 April 20, producing a clear transit event on three sepa-
224
+ rate telescopes and securing that period solution. A further full
225
+ transit was obtained during the next transit window. We describe
226
+ each of these observations in the following section. The data
227
+ themselves will be shown in Sect. 4.
228
+
229
+ 2.2.1. LCOGT
230
+
231
+ We observed TOI-2257 in Sloan i(cid:48) on UTC 2021 March 4
232
+ from the Las Cumbres Observatory Global Telescope (LCOGT;
233
+ Brown et al. 2013) 1.0-meter network node at McDonald Obser-
234
+ vatory. We used the TESS Transit Finder, which is a cus-
235
+ tomized version of the Tapir software package (Jensen 2013),
236
+ to schedule our transit observation. The 4096
237
+ 4096 LCOGT
238
+ Sinistro cameras have an image scale of 0.(cid:48)(cid:48)389 per pixel, result-
239
+ ing in a 26(cid:48) ×
240
+ 26(cid:48) field of view. The images were calibrated via
241
+ the standard LCOGT BANZAI pipeline (McCully et al. 2018),
242
+ and photometric data were extracted with AstroImageJ (Collins
243
+ et al. 2017). The observation covered the 58.6-day orbital period
244
+
245
+ ×
246
+
247
+ N. Schanche et al.: TOI-2257 b
248
+
249
+ Fig. 2. TESS target pixel files
250
+ of Sectors 14, 20, 21, and 26
251
+ that observed TOI-2257, generated
252
+ by means of tpfplotter (Aller
253
+ et al. 2020). The apertures used
254
+ to extract the photometry by the
255
+ SPOC pipeline are shown as red
256
+ shaded regions. The Gaia DR2 cat-
257
+ alog (Gaia Collaboration 2018) is
258
+ over-plotted, with all sources of up
259
+ to 6 magnitudes in contrast with
260
+ TOI-2257 shown as red circles. We
261
+ note that the symbol size scales
262
+ with the magnitude contrast. While
263
+ the star is relatively isolated, there
264
+ is a small amount of contamination
265
+ from outside sources, ranging from
266
+ 2–5% of the total flux.
267
+
268
+ later during the next transit window on 2021 May 25 with 402
269
+ observations using 25s cadence. Due to the low viewing angle,
270
+ the telescope was reoriented shortly before egress, resulting in a
271
+ slight reduction in precision in the final 84 images. The data were
272
+ reduced using the custom pipeline Photometric Reduction and
273
+ In-depth Nightly Curve Exploration (PRINCE; Demory et al.
274
+ 2020).
275
+
276
+ 2.2.4. FLWO
277
+
278
+ We observed TOI-2257 on UTC 2021 April 20 from the Fred
279
+ Lawrence Whipple Observatory (FLWO) on Mt. Hopkins, Ari-
280
+ zona, USA, using the KeplerCam CCD on the 1.2-meter tele-
281
+ 2 binning
282
+ scope. KeplerCam is a 4096
283
+ ×
284
+ 23.1(cid:48)
285
+ mode that produces a 0.(cid:48)(cid:48)672 per pixel scale and a 23.1(cid:48) ×
286
+ field of view. The images were reduced using standard IDL rou-
287
+ tines, and AstroImageJ was used to perform aperture photometry.
288
+ We obtained 245 observations in the i(cid:48) band over a period of 302
289
+ min, resulting in a clear egress detection on target.
290
+
291
+ 4096 detector used in 2
292
+
293
+ ×
294
+
295
+ Further observations were taken on UTC 2021 May 25 using
296
+ the KeplerCam CCD in alternating z(cid:48) and r(cid:48) filters for a total of
297
+ 44 measurements in each filter with exposure times of 90 and
298
+ 120 s, respectively.
299
+
300
+ 2.2.5. SPECULOOS North
301
+
302
+ In addition to observations of the transit events, we also obtained
303
+ 93.5 h of observations to monitor TOI-2257 from the tele-
304
+ scopes associated with the SPECULOOS group and located in
305
+ the Northern Hemisphere; these are the SAINT-EX telescope
306
+ as described in Sect. 2.2.3 and Artemis at the SPECULOOS
307
+ Northern Observatory (SNO) at Teide Observatory in Tenerife
308
+ (Delrez et al. 2018). The SNO data were reduced using the prose
309
+ (Garcia et al. 2021) Python framework, and the SAINT-EX data
310
+
311
+ A45, page 3 of 18
312
+
313
+ alias. The expected transit was ruled out, thus eliminating this
314
+ orbital period from the list of aliases allowed by the TESS data.
315
+
316
+ 2.2.2. TRAPPIST-North
317
+
318
+ We observed TOI-2257 on UTC 2021 April 20 from the
319
+ TRAPPIST-North telescope, located at the Oukaïmeden obser-
320
+ vatory in Morocco (Jehin et al. 2011; Gillon et al. 2011; Barkaoui
321
+ et al. 2019). TRAPPIST-North is a 0.6-meter Ritchey-Chrétien
322
+ telescope equipped with an Andor charge-coupled device (CCD)
323
+ 20(cid:48) with 0.60(cid:48)(cid:48)
324
+ camera that provides a field of view of 20(cid:48) ×
325
+ pixels. This observation consisted of 369 frames of 60s expo-
326
+ sure in the z(cid:48) filter and captured the ingress of the TOI-2257
327
+ planetary candidate transit. The data were reduced using prose
328
+ (Garcia et al. 2021), including frame alignment, calibration, and
329
+ photometric extraction.
330
+
331
+ 2.2.3. SAINT-EX
332
+
333
+ We used the Search And characterIsatioN of Transiting EXo-
334
+ planets (SAINT-EX) Observatory (Demory et al. 2020) to obtain
335
+ one partial and one full transit of TOI-2257. SAINT-EX is a
336
+ 1-meter F/8 Ritchey-Chrétien telescope located at the Observa-
337
+ torio Astronómico Nacional in the Sierra de San Pedro Mártir,
338
+ Mexico. SAINT-EX has an Andor iKon-L camera with deep-
339
+ depletion e2v 2k
340
+ 2k CCD optimized in the near infrared (NIR).
341
+ ×
342
+ The detector field of view is 12(cid:48) ×
343
+ The observations were made with an I+z filter (transmit-
344
+ tance >90% from 750 nm to beyond 1000 nm) designed for the
345
+ observation of faint red targets usually observed by the SPECU-
346
+ LOOS survey (Delrez et al. 2018; Murray et al. 2020; Sebastian
347
+ et al. 2021). The first observation sequence on 2021 April 20
348
+ contained 511 observations with 20s cadence and began shortly
349
+ after the expected ingress. A full transit was observed 35 days
350
+
351
+ 12(cid:48) with 0.34(cid:48)(cid:48) per pixel.
352
+
353
+ A&A 657, A45 (2022)
354
+
355
+ were reduced using the custom PRINCE pipeline. The results of
356
+ this monitoring are provided in Sect. 6.2.
357
+
358
+ 2.3. Spectroscopy
359
+
360
+ In order to better constrain the stellar properties, we also
361
+ obtained two spectra. The results of the spectral analysis are
362
+ included in Sect. 3.
363
+
364
+ 2.3.1. NOT/ALFOSC
365
+
366
+ TOI-2257 was observed with the Alhambra Faint Object Spec-
367
+ trograph and Camera (ALFOSC) mounted on the 2.5-m Nordic
368
+ Optical Telescope (NOT) at the Observatorio Roque de los
369
+ Muchachos (ORM) in La Palma on 2021 June 17 in long slit
370
+ spectroscopy mode. The grism used was gr5, which covers the
371
+ wavelength range between 5000 and 10 700 Å, and the slit width
372
+ was 1.(cid:48)(cid:48)3, leading to a resolution of
373
+ 320. Two exposures of 90 s
374
+ were obtained. The spectroscopic standard star SP1446+259 was
375
+ also observed to correct for instrumental response. Some bias,
376
+ flat field, and arc images were also acquired for the reduction
377
+ and calibration of the spectra.
378
+
379
+ ∼
380
+
381
+ The reduction was performed using IRAF standard routines,
382
+ which included bias subtraction, flat field correction, wavelength
383
+ calibration, and extraction of the spectra. The instrumental
384
+ response was corrected using the spectrum of the spectroscopic
385
+ standard star. The spectra were not corrected for atmospheric
386
+ telluric absorptions.
387
+
388
+ 2.3.2. DBSP
389
+
390
+ −
391
+
392
+ ∼
393
+
394
+ 1.(cid:48)(cid:48)5
395
+
396
+ We observed TOI-2257 on UTC 2021 June 16 from the Hale
397
+ 200(cid:48)(cid:48) telescope at Palomar Observatory using the dual-beam
398
+ optical Double Spectrograph (DBSP). The observations suffered
399
+ from variable cloud cover, and seeing at the start of the night
400
+ 2.(cid:48)(cid:48)0. We obtained two 450 s observations using the
401
+ was
402
+ 1 blue grism (λblaze =
403
+ 1.(cid:48)(cid:48)5 slit, the 5600Å dichroic, the 600 (cid:96) mm−
404
+ 1 red grating (λblaze = 8500 Å).
405
+ 4000 Å), and the 400 (cid:96) mm−
406
+ This instrument configuration covers the full optical window at
407
+ moderate resolving power, R
408
+ 1000, with a modest gap at the
409
+ dichroic. Relative flux calibration was obtained using observa-
410
+ tions of the subdwarf O star HZ44 from Massey & Gronwall
411
+ (1990) obtained on the same night. The spectra were reduced
412
+ using standard IRAF routines.
413
+
414
+ ≈
415
+
416
+ 2.4. High-resolution imaging
417
+
418
+ Two high angular resolution images were taken as part of the
419
+ TESS Follow-up Observing Program (TFOP) in order to rule out
420
+ false positives caused by possible unresolved stellar companions
421
+ or detect any close companions that could dilute the transit sig-
422
+ nal and lead to an underestimation of the planet radius. These
423
+ observations are presented below.
424
+
425
+ 2.4.1. SPP speckle interferometry
426
+
427
+ TOI-2257 was observed using speckle interferometry on 2020
428
+ November 29 with the SPeckle Polarimeter (SPP; Safonov et al.
429
+ 2017) on the 2.5-m telescope at the Sternberg Astronomical Insti-
430
+ tute of Lomonosov Moscow State University (SAI MSU). The
431
+ SPP uses Electron Multiplying CCD Andor iXon 897 as a detec-
432
+ 1. The observation was
433
+ tor with a pixel scale of 20.6 mas px−
434
+ made in the I band, and the atmospheric dispersion was com-
435
+ pensated for. The angular resolution was 89 mas. The detection
436
+
437
+ A45, page 4 of 18
438
+
439
+ Fig. 3. Contrast curves for TOI-2257 b from high angular resolu-
440
+ tion imaging. The top plot shows the results of the Alopeke speckle
441
+ instrument on Gemini North. The bottom plot shows the speckle inter-
442
+ ferometry taken by SPP at the Sternberg Astronomical Institute. Both
443
+ plots have the final reconstructed image inlaid in the upper right. Nei-
444
+ ther instrument detects a nearby companion with a magnitude within
445
+ 4.5 mags of the target star.
446
+
447
+ limit for faint stellar companions is provided in Fig. 3. We did
448
+ not detect any companion brighter than this limit (e.g., 4.5 mag
449
+ at 1(cid:48)(cid:48)).
450
+
451
+ 2.4.2. Alopeke speckle imaging
452
+
453
+ TOI-2257 was observed on UT 2021 February 2 using the
454
+ ‘Alopeke speckle instrument on Gemini North2 (Scott et al.
455
+ 2021). ‘Alopeke provides simultaneous speckle imaging in two
456
+ bands (562 and 832 nm) with output data products that include
457
+ a reconstructed image with robust contrast limits on compan-
458
+ ion detections (e.g., Howell et al. 2016). Eight sets of 1000 X
459
+ 0.06 sec exposures were collected and subjected to Fourier anal-
460
+ ysis in our standard reduction pipeline (see Howell et al. 2011).
461
+ Figure 3 shows our final contrast curves and the 832 nm recon-
462
+ structed speckle image. We find that TOI-2257 is a single star
463
+ with no companion brighter than about 4.5–5 magnitudes below
464
+
465
+ 2 https://www.gemini.edu/sciops/instruments/
466
+ alopeke-zorro/
467
+
468
+ N. Schanche et al.: TOI-2257 b
469
+
470
+ that of the target star from the diffraction limit (20 mas) out to
471
+ 1.2(cid:48)(cid:48). At the distance of TOI-2257 (d = 57.8 pc), these angular
472
+ limits correspond to spatial limits of 1–69 au.
473
+
474
+ 3. Stellar properties
475
+
476
+ 3.1. Spectroscopic analysis
477
+
478
+ Based on the Palomar/DBSP and NOT/ALFOSC spectra, we
479
+ constrained the fundamental properties of TOI-2257 from the
480
+ analysis of molecular band indices.
481
+
482
+ ±
483
+
484
+ ±
485
+
486
+ J).
487
+
488
+ 250 K (CaH2 = 0.36) and Teff = 3395
489
+
490
+ We estimated the effective temperature, Teff, employing the
491
+ relationship between this parameter and the CaH2 index obtained
492
+ for M dwarfs derived by Woolf & Wallerstein (2006). We
493
+ obtained Teff = 3277
494
+ ±
495
+ 310 K (CaH2 = 0.43) for the NOT/ALFOSC and Palomar/DBSP
496
+ data, respectively. These values are in good agreement with those
497
+ obtained from the optical and infrared photometric calibration of
498
+ Casagrande et al. (2008), which yields Teff = 3279
499
+ 51 K for
500
+ (V
501
+
502
+ Ks) and Teff = 3266
503
+ 30 K for (V
504
+ −
505
+ Based on the band strength indices CaH2, CaH3, and TiO5,
506
+ we computed the coarse metallicity parameter ζTiO/CaH (ζ for
507
+ short), as described by Lépine et al. (2007). Then, we used
508
+ the relation between this parameter and the [Fe/H] derived by
509
+ 0.31 (ζ = 0.76
510
+ Mann et al. (2013) to obtain [Fe/H] =
511
+ 0.3) for
512
+ ±
513
+ ±
514
+ the NOT/ALFOSC and Palomar/DBSP data, respectively. These
515
+ values are in good agreement with the results obtained from
516
+ photometric calibrations. We derived [Fe/H]=
517
+ 0.18 dex
518
+ Ks) calibration of Schlaufman & Laughlin (2010),
519
+ for the (V
520
+ −
521
+ Ks) calibration of Mann
522
+ [Fe/H] =
523
+ 0.23
524
+ −
525
+ Ks)
526
+ et al. (2013), and [Fe/H] =
527
+ calibration of Rains et al. (2021).
528
+
529
+ 0.29
530
+ 0.5 dex (ζ = 0.76
531
+
532
+ −
533
+ 0.20 dex for the (Bp
534
+
535
+ 0.22 dex for the (V
536
+
537
+ 0.3) and [Fe/H] =
538
+
539
+ 0.35
540
+
541
+ 0.23
542
+
543
+ 0.23
544
+
545
+ −
546
+
547
+ ±
548
+
549
+ −
550
+
551
+ −
552
+
553
+ ±
554
+
555
+ −
556
+
557
+ ±
558
+
559
+ −
560
+
561
+ −
562
+
563
+ ±
564
+
565
+ ±
566
+
567
+ ±
568
+
569
+ We derived several gravity-sensitive spectral indices to estab-
570
+ lish the main-sequence dwarf nature of TOI-2257. We computed
571
+ the gravity-sensitive indices Na8189 and TiO7140, which, as shown
572
+ by several authors, clearly separate low, intermediate, and high
573
+ gravity for spectral types later than M 2 (e.g., Slesnick et al.
574
+ 2006). The measured spectral indices for the NOT/ALFOSC
575
+ data are TiO7140 = 1.40 and Na8189 = 0.94, whilst for the
576
+ Palomar/DBSP we obtained TiO7140= 1.38 and Na8189 = 0.91.
577
+ These values place TOI-2257 in the high surface gravity region
578
+ (dwarf stars) around the M3 spectral types in Fig. 11 of Slesnick
579
+ et al. (2006). As expected, a visual inspection of the spectra of
580
+ TOI-2257 reveals a clear absorption by the NaI doublet at 8183
581
+ and 8195 Å, which is not seen in giant stars. Furthermore, as
582
+ in Alonso-Floriano et al. (2015), we computed the index Ratio
583
+ C, which is also highly sensitive to surface gravity. We derived
584
+ Ratio C = 1.15 and Ratio C = 1.23 for the NOT/ALFOSC and
585
+ Palomar/DBSP data, respectively. These values place TOI-2257
586
+ in the high surface gravity field-dwarf region in Fig. 6 of Alonso-
587
+ Floriano et al. (2015), which is well above the limit of giant stars
588
+ at Ratio C = 1.07.
589
+
590
+ Finally, we assessed the spectral type of TOI-2257 using the
591
+ PYHAMMER tool3 (Kesseli et al. 2017), which estimates MK
592
+ spectral types by comparing our data with reference spectra of
593
+ M-type stars (Covey et al. 2007). The best fit is obtained for an
594
+ M3V star with [Fe/H]
595
+ 0 dex. Moreover, as in Slesnick et al.
596
+ (2006), we computed the temperature-sensitive index TiO8465,
597
+ in addition to the TiO7140 index presented above, obtaining
598
+ TiO8465 = 1.16 and TiO8465 = 1.07 for the NOT/ALFOSC and
599
+ Palomar/DBSP data, respectively. These values locate TOI-2257
600
+
601
+ ∼
602
+
603
+ 3 https://github.com/BU-hammerTeam/PyHammer
604
+
605
+ Fig. 4. SED of TOI-2257. The red symbols represent the observed pho-
606
+ tometric measurements; the horizontal bars represent the effective width
607
+ of the passband. The blue symbols are the model fluxes from the best-fit
608
+ NextGen atmosphere model (black).
609
+
610
+ around the M3 spectral type region in Fig. 9 of Slesnick et al.
611
+ (2006).
612
+
613
+ 3.2. SED fitting and evolutionary modeling
614
+
615
+ As an independent determination of the basic stellar parameters,
616
+ we performed an analysis of the broadband spectral energy dis-
617
+ tribution (SED) of the star together with the Gaia Early Data
618
+ Release 3 (EDR3; Stassun & Torres 2021) parallax (with no sys-
619
+ tematic correction) in order to determine an empirical measure-
620
+ ment of the stellar radius, following the procedures described
621
+ in Stassun & Torres (2016); Stassun et al. (2017, 2018a). We
622
+ pulled the JHKS magnitudes from 2MASS, the W1–W3 mag-
623
+ nitudes from WISE, the GGBPGRP magnitudes from Gaia, and
624
+ the gry magnitudes from Pan-STARRS. Together, the available
625
+ photometry spans the full stellar SED over the wavelength range
626
+ 0.4–10 µm (see Fig. 4).
627
+
628
+ We performed a fit using NextGen stellar atmosphere mod-
629
+ els (Hauschildt et al. 1999), with the effective temperature (Teff),
630
+ surface gravity (log g), and metallicity ([Fe/H]) adopted from
631
+ the spectroscopic analysis. The remaining free parameter is the
632
+ extinction, AV. The resulting fit (Fig. 4) has a reduced χ2 of 1.6,
633
+ with best fit AV = 0.03
634
+ 0.03. Integrating the (un-reddened)
635
+ ±
636
+ model SED gives the bolometric flux at Earth, Fbol = 1.168
637
+ ±
638
+ 2. Taking the Fbol and Teff together with
639
+ 1 cm−
640
+ 0.041
641
+ 0.015 R
642
+ the Gaia parallax gives the stellar radius R
643
+ .
644
+ ∗
645
+ (cid:12)
646
+ Finally, the absolute MK magnitude together with the empirical
647
+ =
648
+ relations of Mann et al. (2019) implies a stellar mass of M
649
+ 0.33
650
+ . Together with the empirical radius above, this
651
+ 3
652
+ in turn yields a mean stellar density of ρ
653
+ 2.5 g cm−
654
+ ∗
655
+ 0.050.
656
+ and a surface gravity of log g = 4.971
657
+
658
+ 10 erg s−
659
+
660
+ = 0.311
661
+
662
+ 0.02 M
663
+
664
+ = 15.7
665
+
666
+ 10−
667
+
668
+ ±
669
+
670
+ ×
671
+
672
+ ±
673
+
674
+ ±
675
+
676
+ (cid:12)
677
+
678
+ ∗
679
+
680
+ We also estimated the stellar properties using a complemen-
681
+ tary, isochrone-dependent approach. For this analysis, we used
682
+ the isochrones software package (Morton 2015), which can
683
+ be used to fit data inputs to the MESA Isochrones and Stellar
684
+ Tracks database (Dotter 2016; Choi et al. 2016) via the nested
685
+ sampling algorithm MULTINEST (Feroz et al. 2009) as imple-
686
+ mented in the PyMultiNest package (Buchner et al. 2014). We
687
+ used as inputs the stellar metallicity from Table 2, the BV magni-
688
+ tudes tabulated in the revised TESS Input Catalog (TIC; Stassun
689
+ et al. 2019), the JHKs magnitudes from 2MASS (Skrutskie et al.
690
+ 2006; Cutri et al. 2003), and the W1–W3 magnitudes from WISE
691
+
692
+ ±
693
+
694
+ A45, page 5 of 18
695
+
696
+ 110λ (µm)-13-12-11-10log λFλ (erg s-1 cm-2) A&A 657, A45 (2022)
697
+
698
+ Table 2. Properties of the host star.
699
+
700
+ Parameter
701
+
702
+ TIC
703
+ 2MASS
704
+ Gaia DR2
705
+ UCAC 4
706
+
707
+ TESS
708
+ B
709
+ V
710
+ g(cid:48)
711
+ r(cid:48)
712
+ i(cid:48)
713
+ Gaia
714
+ J
715
+ H
716
+ K
717
+ WISE 3.4 µm
718
+ WISE 4.6 µm
719
+ WISE 12 µm
720
+ WISE 22 µm
721
+
722
+ 1
723
+
724
+ RA (J2000)
725
+ Dec (J2000)
726
+ pm (RA) mas yr−
727
+ 1
728
+ pm (Dec) mas yr−
729
+ Parallax mas
730
+ Distance pc
731
+ Spectral type
732
+ Teff/ K
733
+ [Fe/H]
734
+ /M
735
+ M
736
+ ∗
737
+ /R
738
+ R
739
+ (cid:12)
740
+ ∗
741
+ log g / dex
742
+ 3
743
+ ρ
744
+ /g cm−
745
+ ∗
746
+ Fbol/erg s−
747
+
748
+ 2
749
+ 1cm−
750
+
751
+ (cid:12)
752
+
753
+ Value
754
+
755
+ Source
756
+
757
+ Designations
758
+ 198 485 881.01
759
+ J12585767+7739416
760
+ 1 716 345 832 872 291 968
761
+ 839-012174
762
+
763
+ Photometric magnitudes
764
+
765
+ 12.9672
766
+ 16.648
767
+ 15.211
768
+ 15.990
769
+ 14.599
770
+ 13.345
771
+
772
+ 14.1615
773
+ 11.47
774
+ 10.89
775
+ 10.673
776
+ 10.514
777
+ 10.355
778
+ 9.972
779
+ 8.601
780
+
781
+ 0.0074
782
+ ±
783
+ 0.098
784
+ ±
785
+ 0.034
786
+ ±
787
+ 0.242
788
+ ±
789
+ 0.057
790
+ ±
791
+ 0.177
792
+ ±
793
+ 0.000437
794
+ ±
795
+ 0.018
796
+ ±
797
+ 0.015
798
+ ±
799
+ 0.016
800
+ ±
801
+ 0.023
802
+ ±
803
+ 0.02
804
+ ±
805
+ 0.039
806
+ ±
807
+ 0.244
808
+ ±
809
+ Stellar properties
810
+ 12:58:57.51
811
+ +77:39:42.18
812
+ 0.021
813
+ 36.035
814
+ −
815
+ ±
816
+ 0.018
817
+ 31.408
818
+ ±
819
+ 0.015
820
+ 17.283
821
+ ±
822
+ 57.7911+0.1053
823
+ 0.1049
824
+ −
825
+ M3
826
+
827
+ 3430
828
+ 0.27
829
+ −
830
+ 0.33
831
+ 0.311
832
+ 4.971
833
+ 15.8
834
+
835
+ 130
836
+ 0.37
837
+ 0.02
838
+ 0.015
839
+ 0.050
840
+ 2.5
841
+
842
+ ±
843
+ ±
844
+ ±
845
+ ±
846
+ ±
847
+ ±
848
+ 0.055)
849
+
850
+ (1.171
851
+
852
+ ±
853
+
854
+ 10
855
+
856
+ 10−
857
+
858
+ ×
859
+
860
+ [1]
861
+ [1]/[2]
862
+ [1]/[2]
863
+ [2]
864
+ [2]
865
+ [2]
866
+ [3]
867
+ [1]/[4]
868
+ [1]/[4]
869
+ [1]/[4]
870
+ [5]
871
+ [5]
872
+ [5]
873
+ [5]
874
+
875
+ [1]
876
+ [1]
877
+ [6]
878
+ [6]
879
+ [6]
880
+ [7]
881
+ [8]
882
+ [8]
883
+ [8]
884
+ [8]
885
+ [8]
886
+ [8]
887
+ [8]
888
+ SED
889
+
890
+ References. 1: TIC (Stassun et al. 2018b), 2: APASS-dr9 (Henden et al.
891
+ 2015), 3: Gaia DR2 (Gaia Collaboration 2018), 4: Catalog of Cool
892
+ Dwarf Targets (Muirhead et al. 2018), 5: WISE (Cutri et al. 2021),
893
+ 6: Gaia DR3 (Gaia Collaboration 2021), 7: Bailer-Jones et al. (2018),
894
+ 8: see Sect. 3.3.
895
+
896
+ ±
897
+
898
+ ±
899
+ = 0.3190
900
+
901
+ ±
902
+ 0.054, a mass of M
903
+ ±
904
+ ∗
905
+ 0.0024, and a density of ρ
906
+ ∗
907
+
908
+ (Wright et al. 2010) as well as the Gaia EDR3 parallax (Gaia
909
+ Collaboration 2016, 2021). This analysis gives a stellar effective
910
+ 19 K, a surface gravity of log g =
911
+ temperature of Teff = 3441
912
+ 0.028 M
913
+ = 0.315
914
+ 4.926
915
+ , a radius of
916
+ (cid:12)
917
+ 3,
918
+ 1.3 g cm−
919
+ = 13.7
920
+ R
921
+ ±
922
+ ∗
923
+ which is in line with the results of the SED fitting. The results
924
+ 8.0 Gyr at 3σ and minimal line-of-
925
+ also favor a stellar age of
926
+ sight extinction (AV < 0.2 at 3σ). While we note the consistency
927
+ of the isochrone-derived parameter values with the others pre-
928
+ sented here, we do not include them in the calculation of the
929
+ final, adopted stellar parameters as the uncertainties are likely
930
+ underestimated.
931
+
932
+ ≥
933
+
934
+ 3.3. Adopted stellar parameters
935
+
936
+ We compiled our final set of adopted stellar parameters from the
937
+ analyses of the NOT/ALFOSC and Palomar/DBSP spectra and
938
+
939
+ A45, page 6 of 18
940
+
941
+ the SED. In particular, the analyses detailed previously yielded
942
+ three estimates of Teff (from NOT/ALFOSC, Palomar/DBSP,
943
+ and the SED analyses) and two estimates of [Fe/H] (from the
944
+ NOT/ALFOSC and Palomar/DBSP analyses). For each param-
945
+ eter, we calculated the weighted mean of the estimates using a
946
+ Monte Carlo approach. We drew a total of 106 samples from
947
+ normal distributions with the means and standard deviations of
948
+ the estimates, using inverse-variance weighting to determine the
949
+ number of subsamples to draw from each distribution. We then
950
+ calculated the mean and standard deviation of the resulting sam-
951
+ ple. We adopted these values as our final estimates of Teff and
952
+ [Fe/H], which are presented in Table 2.
953
+
954
+ 4. Transit analysis
955
+
956
+ In this section we outline the fit of the data to a transit model as
957
+ well as an analysis of possible transit timing variations (TTVs).
958
+
959
+ 4.1. Transit model fit
960
+
961
+ All the light curves described in Sect. 2 were used simulta-
962
+ neously to fit the transit using a Markov chain Monte Carlo
963
+ (MCMC) approach. We employed the PyTransit (Parviainen
964
+ 2015) implementation of the Mandel & Agol (2002) quadratic
965
+ limb darkening transit model, with sampling done using emcee
966
+ (Foreman-Mackey et al. 2013), a Python implementation of the
967
+ affine invariant MCMC ensemble sampler (Goodman & Weare
968
+ 2010). The parameters used for the transit model are the planet-
969
+ to-star radius ratio Rp/R
970
+ , the impact parameter b, the zero epoch
971
+ ∗
972
+ T0, the period P, and the stellar mass and radius M
973
+ . The
974
+ eccentricity and argument of periastron were fit using the param-
975
+ eterization of √e cos ω and √e sin ω. In addition, two quadratic
976
+ limb darkening parameters were fit for each wavelength. The
977
+ ground-based data were simultaneously detrended using an out-
978
+ of-transit baseline as well as quadratic terms for the airmass and
979
+ full width at half maximum (FWHM) trends. Priors used for the
980
+ analysis are shown in Table 3.
981
+
982
+ and R
983
+ ∗
984
+
985
+ ∗
986
+
987
+ The final transit model is shown in Fig. 5 over the detrended
988
+ light curves. The posterior distributions for the transit parame-
989
+ ters can be found in Appendix A.1, and the final fit and derived
990
+ parameters for the planetary system can be found in Table 3.
991
+ While we currently have no mass estimations from RV measure-
992
+ ments, we used the relationship in Chen & Kipping (2017) to
993
+ estimate a range of likely values. The RV semi-amplitude could
994
+ 1.
995
+ thus be estimated, leading to an expected signal around 3.5 m s−
996
+ We also carried out an independent analysis using the
997
+ juliet package (Espinoza et al. 2019), which is built over
998
+ batman (Kreidberg 2015) for the modeling of transits and the
999
+ dynesty (Speagle 2020) dynamic nested sampling algorithm
1000
+ for estimating Bayesian posteriors and evidence. The fit tran-
1001
+ sit parameters were: Rp/R
1002
+ , b, T0, P, the stellar density ρ
1003
+ ,
1004
+ ∗
1005
+ √e cos ω, and √e sin ω. For each passband, we also fit two
1006
+ quadratic limb-darkening coefficients, which were parameter-
1007
+ ized using the (q1, q2) triangular sampling scheme of Kipping
1008
+ (2013). All these parameters were sampled from wide uniform
1009
+ priors, except the stellar density, for which we used a normal
1010
+ prior based on the value and uncertainty reported in Table 2.
1011
+
1012
+ ∗
1013
+
1014
+ We first performed individual analyses of each of our light
1015
+ curves, in order to select for each of them the best correlated
1016
+ noise model based on Bayesian evidence. We explored a large
1017
+ range of models, consisting of first- to fourth-order polynomials
1018
+ with respect to, for example, time, airmass, point spread func-
1019
+ tion (PSF) FWHM, background, stellar position on the detector,
1020
+
1021
+ Table 3. Fit and derived parameters for the TOI-2257 b system.
1022
+
1023
+ N. Schanche et al.: TOI-2257 b
1024
+
1025
+ Parameter
1026
+
1027
+ Unit
1028
+
1029
+ Value
1030
+
1031
+ Fitted parameters
1032
+
1033
+ Prior
1034
+
1035
+ ∗
1036
+
1037
+ )
1038
+
1039
+ Orbital period P
1040
+ Mid-transit time T0
1041
+ (Rp/R
1042
+ Impact parameter b
1043
+ √e sin ω
1044
+ √e cos ω
1045
+ Stellar mass M
1046
+ ∗
1047
+ Stellar radius R
1048
+ ∗
1049
+
1050
+ days
1051
+ BJD-2 450 000
1052
+
1053
+ M
1054
+ (cid:12)
1055
+ R
1056
+
1057
+ (cid:12)
1058
+
1059
+ N
1060
+ 0.00105 N
1061
+
1062
+ 35.189346(90)
1063
+ 9007.97949+0.00108
1064
+ −
1065
+ 0.06423+0.00142
1066
+ 0.00133
1067
+ −
1068
+ 0.374+0.098
1069
+ 0.137
1070
+ −
1071
+ 0.615+0.083
1072
+ 0.073
1073
+ −
1074
+ −
1075
+ 0.126+0.484
1076
+ 0.387
1077
+ −
1078
+ −
1079
+ 0.328+0.021
1080
+ 0.019
1081
+ −
1082
+ 0.015
1083
+ ±
1084
+
1085
+ 0.313
1086
+
1087
+ −
1088
+
1089
+ U
1090
+
1091
+ (35.189295, 1e
1092
+ 4)
1093
+ (9007.978906, 0.1)
1094
+ (0.001, 0.4)
1095
+ (0, 1)
1096
+ U
1097
+ 1, 1)
1098
+ (
1099
+ −
1100
+ 1, 1)
1101
+ (
1102
+ −
1103
+ (0.33, 0.02)
1104
+ N
1105
+ (0.311, 0.015)
1106
+
1107
+ U
1108
+ U
1109
+
1110
+ N
1111
+
1112
+ Physical and orbital parameters
1113
+
1114
+ Planet radius Rp
1115
+ Orbital eccentricity e
1116
+ Argument of periastron ω
1117
+ Semimajor axis a
1118
+ Inclination i
1119
+ Equilibrium temperature Teq
1120
+ Depth δ
1121
+ Transit duration
1122
+
1123
+ (
1124
+
1125
+ )
1126
+ ∗
1127
+
1128
+ R
1129
+
1130
+ ⊕
1131
+
1132
+ ◦
1133
+ au
1134
+ ◦
1135
+ K
1136
+
1137
+ h
1138
+
1139
+ Planet mass Mp
1140
+ RV Semi-amplitude K
1141
+ TSM
1142
+
1143
+ M
1144
+ ⊕
1145
+ 1
1146
+ m s−
1147
+
1148
+ Predicted parameters
1149
+
1150
+ Limb darkening
1151
+
1152
+ 2.194+0.113
1153
+ 0.111
1154
+ −
1155
+ 0.496+0.216
1156
+ 0.133
1157
+ −
1158
+ 101.674+42.453
1159
+ 26.881
1160
+ −
1161
+ −
1162
+ 0.145
1163
+ 0.003
1164
+ ±
1165
+ 89.786+0.078
1166
+ 0.062
1167
+ −
1168
+ 256+61
1169
+ 17
1170
+ −
1171
+ 0.00413+0.00018
1172
+ 0.00017
1173
+ −
1174
+ 3.846+0.057
1175
+ 0.051
1176
+ −
1177
+
1178
+ 5.712+4.288
1179
+ 2.311
1180
+ −
1181
+ 3.521+2.901
1182
+ 1.507
1183
+ −
1184
+ 32.708+24.08
1185
+ 14.362
1186
+ −
1187
+
1188
+ q1 (TESS)
1189
+ q2 (TESS)
1190
+ q1 (I + z)
1191
+ q2 (I + z)
1192
+ q1 (i(cid:48))
1193
+ q2 (i(cid:48))
1194
+ q1 (z(cid:48))
1195
+ q2 (z(cid:48))
1196
+ q1 (r(cid:48))
1197
+ q2 (r(cid:48))
1198
+
1199
+ 0.328
1200
+ 0.203
1201
+ 0.289
1202
+
1203
+ 0.008
1204
+ 0.022
1205
+ 0.008
1206
+
1207
+ ±
1208
+ ±
1209
+ ±
1210
+ 0.193+0.027
1211
+ 0.026
1212
+ −
1213
+ 0.385+0.01
1214
+ 0.009
1215
+ −
1216
+ 0.03
1217
+ 0.213
1218
+ ±
1219
+ 0.300+0.008
1220
+ 0.009
1221
+ −
1222
+ 0.206+0.034
1223
+ 0.031
1224
+ −
1225
+ 0.568+0.012
1226
+ 0.011
1227
+ −
1228
+ 0.034
1229
+ ±
1230
+
1231
+ 0.164
1232
+
1233
+ (0.3281, 0.0079)
1234
+ (0.2020, 0.0229)
1235
+ (0.2903, 0.0076)
1236
+ (0.1992, 0.0260)
1237
+ (0.3849, 0.0098)
1238
+ (0.2070, 0.0301)
1239
+ (0.2994, 0.0088)
1240
+ (0.2050, 0.0327)
1241
+ (0.5686, 0.0116)
1242
+ (0.1663, 0.0326)
1243
+
1244
+ N
1245
+ N
1246
+ N
1247
+ N
1248
+ N
1249
+ N
1250
+ N
1251
+ N
1252
+ N
1253
+ N
1254
+
1255
+ )Time-averaged equilibrium temperature taking into account the eccentric orbit using the formulation of Méndez & Rivera-Valentín (2017),
1256
+ ∗
1257
+
1258
+ Notes. (
1259
+ assuming zero albedo and full day-night heat redistribution.
1260
+
1261
+ or any combination of these parameters. First- or second-order
1262
+ polynomials of airmass and/or FWHM were typically favored.
1263
+ For each light curve, we also fit a jitter term, which was added
1264
+ quadratically to the error bars of the data points, to account for
1265
+ any underestimation of the uncertainties or any excess noise not
1266
+ captured by our modeling.
1267
+
1268
+ We then conducted two global analyses: one assuming a cir-
1269
+ cular orbit (e set to zero) and one allowing the orbit of the planet
1270
+ to be eccentric. Following Espinoza et al. (2019), we found the
1271
+ eccentric model (e = 0.41+0.21
1272
+ 0.12) to be strongly favored over the
1273
+ −
1274
+ circular one, with a difference in Bayesian log evidence (∆ ln Z)
1275
+ 180 : 1 assuming equiprobable
1276
+ of 5.2 (i.e., posterior odds of
1277
+
1278
+ ≈
1279
+
1280
+ models), lending substantial support to our eccentric interpreta-
1281
+ tion (Kass & Raftery 1995). All transit parameters returned by
1282
+ the eccentric fit are consistent within the uncertainties with the
1283
+ ones reported in Table 3.
1284
+
1285
+ 4.2. Transiting timing variations
1286
+
1287
+ Transit timing variations are of interest in this context because
1288
+ they can point to the existence of other planets in the system and
1289
+ can potentially be used to constrain the mass of the transiting
1290
+ planet. We investigated the evidence for TTVs of TOI-2257 b
1291
+ in a separate analysis of the transit light curves, generating the
1292
+
1293
+ A45, page 7 of 18
1294
+
1295
+ A&A 657, A45 (2022)
1296
+
1297
+ Fig. 5. Photometric observations of TOI-2257 with the best-fit transit parameters over-plotted. Thick black points show 10-min bins, with all data
1298
+ points shown in gray. Residuals from the model fit are shown with an arbitrary offset below the transit. It should be noted that, for the sake of
1299
+ clarity, only 10-min-binned residuals are shown. The color of the line reflects the filter used for the observations (red for TESS, orange for I+z,
1300
+ green for i’, blue for z’, and purple for r’).
1301
+
1302
+ A45, page 8 of 18
1303
+
1304
+ 0.970.980.991.001.011.02TESS sector 140.970.980.991.001.011.02TESS sector 200.970.980.991.001.011.02SAINT-EX (April)0.970.980.991.001.011.02SAINT-EX (May)0.970.980.991.001.011.02KeplerCam i'0.970.980.991.001.011.02KeplerCam z'32101230.970.980.991.001.011.02TRAPPIST-N32101230.970.980.991.001.011.02KeplerCam r'Time from mid-transit (hours)Normalised flux N. Schanche et al.: TOI-2257 b
1305
+
1306
+ Fig. 6. Archival images around TOI-2257 with TESS’s apertures used in Sectors 14 (red), 20 (blue), 21 (yellow) and 26 (magenta) superimposed
1307
+ to assess for current, unresolved blending. From left to right: (1) 1955 February 28 DSS1-red filter, (2) 1998 March 23 DSS2-red filter, (3) 2013
1308
+ September 06 PTF-red filter, and (4) 2021 April 20 TN-red filter. Its current location (red cross) is marked in all images.
1309
+
1310
+ TTV light curve model with the exoplanet software package
1311
+ (Foreman-Mackey et al. 2021a) and performing the posterior
1312
+ inference with the PyMC3 software package (Salvatier et al.
1313
+ 2016). Our approach was similar to those of other recent TTV
1314
+ analyses of TESS-discovered exoplanets (Badenas-Agusti et al.
1315
+ 2020; Daylan et al. 2021).
1316
+
1317
+ As data inputs, we used TESS Sector 14 and 20 two-minute-
1318
+ cadence light curves, the 2021 April 20 partial-transit light
1319
+ curves from SAINT-EX, KeplerCam, and TRAPPIST-N, and the
1320
+ 2021 May 25 full-transit light curve from SAINT-EX, the most
1321
+ constraining data on this particular transit. We placed normal
1322
+ priors on the stellar mass and radius using the values presented
1323
+ in Table 2 and on the planet-to-star radius ratio and period using
1324
+ the values presented in Table 3. We placed uninformative priors
1325
+ on the impact parameter and the quadratic limb-darkening coef-
1326
+ ficients using the physical distributions built into exoplanet,
1327
+ the latter of which relies on the triangular sampling method of
1328
+ Kipping (2013). We placed a normal prior on the period of the
1329
+ planet using the results of the linear-ephemeris analysis and uni-
1330
+ form priors on the times of the four observed transits, centered
1331
+ on the expected transit times from the linear ephemeris and with
1332
+ a width of 1 h. Finally, we placed a normal prior on the out-of-
1333
+ transit flux for all data sets, centered on the median flux value
1334
+ and using a standard deviation that is robustly estimated using
1335
+ the median absolute deviation of the fluxes.
1336
+
1337
+ This model has some notable simplifications. We assumed
1338
+ that the noise is normally distributed and did not attempt to
1339
+ model any residual systematics in the normalized light curves.
1340
+ Since all the observations are generally in the red optical, we
1341
+ also used a single set of limb-darkening coefficients and a single
1342
+ transit depth to model the light curve despite the somewhat dif-
1343
+ ferent bandpasses of the observations. Finally, we used a single
1344
+ out-of-transit flux value for the combined data set, assuming that
1345
+ any errors in the normalization of the individual data sets are
1346
+ negligible. These simplifications notably reduce the complexity
1347
+ of the model and improve the computational efficiency of the
1348
+ sampling, and we assume that they do not significantly impact
1349
+ the inference of the transit times.
1350
+
1351
+ ±
1352
+
1353
+ ±
1354
+
1355
+ 2.0
1356
+
1357
+ 6.4, 0.3
1358
+
1359
+ 1.1, and 0.3
1360
+
1361
+ ephemeris at the 1σ level with inferred TTVs of 1.5
1362
+
1363
+ We find that the four transits are consistent with a linear
1364
+ 4.7,
1365
+ 1.2 min. The smaller uncertainties
1366
+ −
1367
+ of the last two transits illustrate the improved timing estimates
1368
+ afforded by our ground-based follow-up observations. We con-
1369
+ clude that the existing observations do not show any evidence
1370
+ of TTVs. TESS observations of TOI-2257 in the near future (in
1371
+ Sectors 40, 41, 47, 48, and 53) can be monitored to check if any
1372
+ potential variations emerge.
1373
+
1374
+ ±
1375
+
1376
+ ±
1377
+
1378
+ 5. Planet validation
1379
+
1380
+ 5.1. TESS data validation report
1381
+
1382
+ The natural first step for false-positive vetting was to closely ana-
1383
+ lyze the TESS data validation report (Twicken et al. 2018; Li
1384
+ et al. 2019) based on the 176-day solution. All tests, including
1385
+ those for odd-even transit depths, centroid shifts, and “ghosts”,
1386
+ produced favorable results. The encouraging results led to fur-
1387
+ ther follow-up in the system.
1388
+
1389
+ 5.2. Follow-up photometry
1390
+
1391
+ The TESS data validation report indicates that there is one
1392
+ contaminating source within the TESS apertures used in both
1393
+ Sectors 14 and 20 at a distance of 16.47(cid:48)(cid:48). We identify this as the
1394
+ faint source with Gaia ID 1716345832871506560 (G-mag 20.7).
1395
+ The pixel scale from the ground-based follow-up photometry
1396
+ extracted light curves using an aperture of only a few arcseconds,
1397
+ and therefore we were able to resolve the target in isolation. No
1398
+ transit event was seen on this faint star or any other nearby star,
1399
+ while a dimming event was clearly observed on the target.
1400
+
1401
+ 5.3. Archival imaging
1402
+
1403
+ ∼
1404
+
1405
+ 0.047 arcsec yr−
1406
+
1407
+ We used archival images to investigate the contamination of
1408
+ background stars (e.g., Quinn et al. 2019). We aimed to rule out
1409
+ the presence of eclipsing binaries (EBs) at the present-day target
1410
+ location, which might introduce transit-like signals in the data.
1411
+ Unfortunately, TOI-2257 has a moderately low proper motion
1412
+ 1. The oldest archival image that we found
1413
+ of
1414
+ was taken in 1955, 64 and 66 yr before the first and the last
1415
+ TESS observations in Sectors 14 and 26, respectively. This oldest
1416
+ 1
1417
+ archival image has a pixel scale and PSF of 1.69 arcsec pixel−
1418
+ and 8.05 arcsec, respectively. Since that time, the star has only
1419
+ moved
1420
+ 3.1 arcsec (see Fig. 6). While it seems that there is not
1421
+ any background star that might be producing the transit-like sig-
1422
+ nals detected in our data, we cannot confidently rule out that
1423
+ 150 yr to reach a separa-
1424
+ possibility. We would need to wait for
1425
+ tion of 10 arcsec since 1955, a distance larger than the 1955 PSF
1426
+ that would allow us to confidently rule out the contamination
1427
+ caused by a background EB.
1428
+
1429
+ ∼
1430
+
1431
+ ∼
1432
+
1433
+ 5.4. False-positive likelihood
1434
+
1435
+ To assess the possibility that the observed transit was not due
1436
+ to a planet orbiting the target star, but rather from an astro-
1437
+ physical false-positive scenario, we used the software package
1438
+
1439
+ A45, page 9 of 18
1440
+
1441
+ A&A 657, A45 (2022)
1442
+
1443
+ triceratops (Giacalone et al. 2021). This package was devel-
1444
+ oped as a tool to assist in the vetting and validation of TESS
1445
+ candidates (see, e.g., Cloutier et al. 2020; Demory et al. 2020;
1446
+ Hedges et al. 2021; Wells et al. 2021). Using a Bayesian frame-
1447
+ work, triceratops calculates the probability that the signal is
1448
+ caused by a variety of true-positive or false-positive scenarios.
1449
+ The calculations incorporate prior knowledge about the target
1450
+ star, exoplanet occurrence rates, and stellar multiplicity. One
1451
+ metric that is returned is the false-positive probability (FPP),
1452
+ which is the sum of probabilities for all false-positive scenar-
1453
+ ios. Using the TESS 2-min data from Sectors 14 and 20 gives
1454
+ 0.009. However, triceratops is also able to
1455
+ an FPP of 0.078
1456
+ incorporate additional information from contrast curves to fur-
1457
+ ther constrain the false-positive scenarios. By incorporating this
1458
+ information, the FPP is reduced to 0.0255
1459
+ 0.0012. However,
1460
+ the higher precision of the SAINT-EX light curve is more con-
1461
+ straining than the TESS data. We therefore modified the input
1462
+ to triceratops to use the detrended SAINT-EX transit obser-
1463
+ vation from May 25 along with the contrast curve, reducing the
1464
+ 9). We are therefore able to
1465
+ FPP to a negligible value (
1466
+ statistically validate this object as a planet.
1467
+
1468
+ 10−
1469
+
1470
+ 8.7
1471
+
1472
+ ×
1473
+
1474
+ ±
1475
+
1476
+ ∼
1477
+
1478
+ ±
1479
+
1480
+ 5.5. Unresolved stellar companions
1481
+
1482
+ Based on the high-resolution imaging (see Sect. 2.4), we ruled
1483
+ out the potential for a companion within 4.5–5.5 mag of the
1484
+ target star outside 6 au at 832 nm, which corresponds to a
1485
+ magnitude of 18.6 in the r(cid:48) band. Theory and observations estab-
1486
+ lished that the mass cutoff for what constitutes a star is different
1487
+ for objects of different metallicity. For objects with a solar-like
1488
+ metallicity, anything with less than 0.075 M
1489
+ will be a brown
1490
+ dwarf (Boss 2001), while for objects with lower metallicity, the
1491
+ mass limit will be about 0.083 M
1492
+ (Richer et al. 2006). Using
1493
+ the models by Baraffe et al. (2015) and assuming a stellar age
1494
+ of 5 Gyr, we found an upper limit for the companion’s mass of
1495
+ 0.075 M
1496
+ . Hence, this result allows us to confidently rule out the
1497
+ (cid:12)
1498
+ presence of an unresolved stellar companion. Nevertheless, sub-
1499
+ stellar objects with masses ranging from 0.01–0.075 M
1500
+ would
1501
+ (cid:12)
1502
+ still be possible.
1503
+
1504
+ (cid:12)
1505
+
1506
+ (cid:12)
1507
+
1508
+ 6. Planet searches and detection limits
1509
+
1510
+ 6.1. TESS photometry
1511
+
1512
+ To search for additional planets, we used our custom pipeline
1513
+ SHERLOCK4 (Pozuelos et al. 2020; Demory et al. 2020).
1514
+ SHERLOCK is a user-friendly open-source package that has five
1515
+ different modules that allow the user to: (1) search for planetary
1516
+ candidates; (2) perform vetting of the most promising signals;
1517
+ (3) compute a statistical validation; (4) model the signals to
1518
+ refine the ephemerides; and (5) compute the observational win-
1519
+ dows from ground-based observatories. SHERLOCK has direct
1520
+ access to short- and long-cadence data observed by Kepler/K2
1521
+ and TESS. Hence, SHERLOCK is a fully operational, power-
1522
+ ful tool that allows users to perform the planet search fast and
1523
+ robustly.
1524
+
1525
+ SHERLOCK applies a multi-detrend approach to the nominal
1526
+ light curve, employing the w¯otan package (Hippke et al. 2019),
1527
+ that is, the nominal light curve is detrended several times using a
1528
+ bi-weight filter by varying the window size. This strategy allows
1529
+
1530
+ the user to maximize the signal detection efficiency (SDE) and
1531
+ the signal-to-noise ratio (S/N) of the transit search, which is
1532
+ performed over the nominal light curve, jointly with the new
1533
+ detrended light curves, through the transit least squares
1534
+ (TLS) package (Hippke & Heller 2019). TLS is optimized to
1535
+ detect shallow periodic transits using an analytical transit model
1536
+ based on the stellar parameters. The transit search is performed
1537
+ in a loop; once a signal is found, it is recorded and masked,
1538
+ and then the search keeps running until no more signals with
1539
+ S /N (cid:62) 5 are found in the data set. To start the search for extra
1540
+ planets, we masked the transits corresponding to the candidate
1541
+ TOI-2257 b, with an orbital period of 35.19 d, T0 = 1691.28 d,
1542
+ and T14 = 228 min. Then, we performed three transit searches,
1543
+ firstly by considering all the sectors available simultaneously,
1544
+ that is, combining Sectors 14, 20, 21, and 22. We focused our
1545
+ search for orbital periods ranging from 0.5 to 30 d, where at
1546
+ least two transits were required to identify a potential signal. Sec-
1547
+ ondly, we focused on the longer orbital periods, ranging from 40
1548
+ to 80 d. In this case, we allowed single transits to be recovered.
1549
+ Finally, we explored all the sectors independently, focusing on
1550
+ orbital periods ranging from 0.5 to 15 d. This strategy allowed
1551
+ us to avoid having sectors with different photometric precision
1552
+ affect the global search.
1553
+
1554
+ After the scrutiny of the data, we found no clear evi-
1555
+ dence of additional planetary transits. All the signals found by
1556
+ SHERLOCK were attributable to systematics, noise, or variabil-
1557
+ ity. Following Wells et al. (2021), the lack of detections of extra
1558
+ signals suggests that: (1) no other planets exist in the system; (2)
1559
+ if they do exist, they do not transit; or (3) they exist and transit,
1560
+ but the photometric precision of the data set is not good enough
1561
+ to detect them, or they have periods longer than the ones explored
1562
+ in this study. If scenario (2) or (3) is true, extra planets might be
1563
+ detected by RV follow-up, as discussed in Sect. 7.
1564
+
1565
+ To evaluate scenario (3), we studied the detection limits
1566
+ of the current data set by performing injection-and-recovery
1567
+ experiments over the PDC-SAP light curves, combining the four
1568
+ sectors available. To this end, we used the MATRIX ToolKit5.
1569
+ MATRIX ToolKit allows the user to define the ranges
1570
+ in the Rplanet–Pplanet parameter space to be examined. Each
1571
+ combination of Rplanet–Pplanet is explored using a number of dif-
1572
+ ferent phases, that is, different values of T0. For simplicity, it
1573
+ is assumed that the impact parameters and eccentricities of the
1574
+ injected planets are zero. To perform the injection-and-recovery
1575
+ experiments, MATRIX ToolKit injected the synthetic planets,
1576
+ detrended the light curves using a bi-weight filter with a window-
1577
+ size of 0.95 d, which was found to be the optimal value during
1578
+ the SHERLOCK runs, and masked the transits corresponding to
1579
+ the candidate TOI-2257 b. We considered a synthetic planet to
1580
+ be recovered when its epoch was detected with 1 h accuracy and
1581
+ the recovered period was within 5% of the injected period. It
1582
+ is worth noting that since we injected the synthetic planets in
1583
+ the PDC-SAP light curve, the signals were not affected by the
1584
+ PDC-SAP systematic corrections; therefore, the detection limits
1585
+ should be considered as the most optimistic scenario (see, e.g.,
1586
+ Pozuelos et al. 2020; Eisner et al. 2020).
1587
+
1588
+ In particular, we explored the Rplanet–Pplanet parameter space
1589
+ in the ranges of 0.5–3.5 R
1590
+ , and 1.0–30.0 d
1591
+ with steps of 1.0 d. For each pair Rplanet–Pplanet we used four dif-
1592
+ ferent phases. Hence, we analyzed a total of 1920 scenarios. The
1593
+ results, shown in Fig. 7, allowed us to rule out planets with sizes
1594
+
1595
+ with steps of 0.2 R
1596
+ ⊕
1597
+
1598
+ ⊕
1599
+
1600
+ 4 The SHERLOCK (Searching for Hints of Exoplanets
1601
+ from
1602
+ Lightcurves of space-based seekers) code is fully available on GitHub:
1603
+ https://github.com/franpoz/SHERLOCK
1604
+
1605
+ 5 The MATRIX ToolKit (Multi-phase Transits Recovery from
1606
+ Injected exoplanets Toolkit) code is available on GitHub: https://
1607
+ github.com/martindevora/tkmatrix
1608
+
1609
+ A45, page 10 of 18
1610
+
1611
+ N. Schanche et al.: TOI-2257 b
1612
+
1613
+ Fig. 7. Injection-and-recovery tests performed on TESS data to check
1614
+ the detectability of extra planets in the system. We explored a total of
1615
+ 1920 different scenarios. Each pixel evaluated four scenarios, that is,
1616
+ four light curves with injected planets having different Pplanet, Rplanet
1617
+ and T0. Larger recovery rates are presented in yellow and green, while
1618
+ lower recovery rates are shown in blue and darker hues. We can rule out
1619
+ the presence of planets with sizes >2.5 R
1620
+ . Planets with sizes <1.5 R
1621
+ ⊕
1622
+ would remain undetected. Planets with sizes between 1.5 and 2.5 R
1623
+ ⊕
1624
+ have recovery rates ranging from 30 to 70%. The red vertical line marks
1625
+ the maximum orbital period (12.54 d) allowed for the hypothetical inner
1626
+ planet to avoid crossing orbits between the planets. See the text for more
1627
+ details.
1628
+
1629
+ ⊕
1630
+
1631
+ >2.5 R
1632
+ , with recovery rates larger than 80% for almost the full
1633
+ ⊕
1634
+ range of periods explored. On the other hand, planets with sizes
1635
+ <1.5 R
1636
+ would remain undetected. In addition, we found that
1637
+ ⊕
1638
+ planet sizes 1.5 < R
1639
+ < 2.5 might be challenging to detect, with
1640
+ ⊕
1641
+ recovery rates ranging from 30 to 70 %.
1642
+
1643
+ It is important to notice that with the current data set, our
1644
+ models for TOI-2257 b presented in Sect. 4 favored an eccentric
1645
+ orbit. As such, there are forbidden periods for the hypotheti-
1646
+ cal inner planet to avoid crossing orbits between the planets.
1647
+ In particular, safe orbits have a semimajor axis lower than the
1648
+ periastron distance for TOI-2257 b (q = 0.073 au). That is, the
1649
+ allowed orbital periods would be (cid:46)12.54 d. This limit in the
1650
+ orbital period for the hypothetical inner planet is displayed in
1651
+ Fig. 7 with a vertical red line.
1652
+
1653
+ 6.2. SAINT-EX + Artemis photometry
1654
+
1655
+ ≤
1656
+
1657
+ 1.5 R
1658
+ ⊕
1659
+
1660
+ As shown previously using TESS photometry, it would be chal-
1661
+ lenging to find small planets with sizes
1662
+ . Accordingly, we
1663
+ monitored TOI-2257 using the SPECULOOS network located
1664
+ in the Northern Hemisphere. This high-precision photometry
1665
+ allows us to detect single transits of Earth-size planets orbit-
1666
+ ing M dwarfs (Delrez et al. 2018; Murray et al. 2020; Demory
1667
+ et al. 2020; Niraula et al. 2020). Outwith the two transits shown
1668
+ in Sect. 2.2.3 we followed up the star for 93.5 h. These time-
1669
+ series observations were analyzed by our automatic pipelines
1670
+ as described in Murray et al. (2020) and Demory et al. (2020)
1671
+ in the search for transit events. We did not find any hint of
1672
+ transit-like features during this time. Hence, this negative result
1673
+ allowed us to rule out the presence of Earth-size planets for a
1674
+ range of orbital periods by computing the phase coverage; that
1675
+ is, we computed the percentage of phase covered for each orbital
1676
+ period from 0.1 to 15 d, in intervals of 0.01 d, for a total of 1500
1677
+ periods. We found that because of day-night cycles only very
1678
+ short periods are fully covered in phase by our observations,
1679
+ as shown in Fig. 8. Indeed, only for orbital periods of
1680
+ 3.78 d
1681
+ was more than 80% of the orbit explored, meaning that periods
1682
+ equal to or shorter than this would most likely be detected if the
1683
+ planet existed and transited. For larger orbital periods, the phase
1684
+
1685
+ ≤
1686
+
1687
+ Fig. 8. Phase coverage for SAINT-EX and Artemis data. Top: evolu-
1688
+ tion of the phase coverage of a hypothetical planet orbiting TOI-2257 as
1689
+ a function of the period, derived from SAINT-EX and Artemis obser-
1690
+ vations (blue dots). The effective phase coverage is the integral of the
1691
+ phase coverage over the 0.1–15 day period range and is equal to 59%.
1692
+ The dotted red (purple) line indicates the period above which the phase
1693
+ coverage is always inferior to 80% (40%), which corresponds to the
1694
+ period of 3.78 days (8.71 days). We note that periods equal to an integer
1695
+ number of days are significantly less covered due to day-night cycles in
1696
+ ground-based observations. Bottom: graphical visualization of the cov-
1697
+ erage of TOI-2257 with SNO and SAINT-EX for a hypothetical planet
1698
+ with orbital periods of 3.78 and 8.71 days, respectively. Each blue circu-
1699
+ lar arc represents one night of observation; its size is proportional to the
1700
+ number of hours observed each night, and a full circle depicts a duration
1701
+ of 3.78 (8.71) days.
1702
+
1703
+ coverage decreases rapidly to a minimum of 20% for the highest
1704
+ orbital period explored, 15 d.
1705
+
1706
+ 7. Discussion
1707
+
1708
+ 7.1. Composition
1709
+
1710
+ 0.111 R
1711
+ ⊕
1712
+ −
1713
+
1714
+ The radius of TOI-2257 b is 2.194+0.113
1715
+ . Not having RV mea-
1716
+ surements, we used the relation of Chen & Kipping (2017) to
1717
+ estimate the likely mass, providing a wide range of predicted
1718
+ values centered around 5.7 M
1719
+ . Using mass-radius curves from
1720
+ Zeng et al. (2016), it is expected that the planet’s density is con-
1721
+ sistent with a composition of an ice or gas giant rather than a
1722
+ denser rocky body.
1723
+
1724
+ ⊕
1725
+
1726
+ 7.2. Eccentricity
1727
+
1728
+ Not having RV measurements, the eccentricity reported here is
1729
+ derived from the transit itself. In particular, it is driven by the
1730
+ transit duration, which can be degenerate with the impact param-
1731
+ eter. Nevertheless, as the eccentric case is statistically favored
1732
+ over the circular case, we explore the system using the result-
1733
+ ing eccentricity of e = 0.496+0.216
1734
+ 0.133. The high eccentricity of
1735
+ −
1736
+ this object, shown in context in Fig. 9, provides a clue to the
1737
+ dynamic history of the system. Single-planet systems tend to
1738
+
1739
+ A45, page 11 of 18
1740
+
1741
+ A&A 657, A45 (2022)
1742
+
1743
+ Fig. 9. Period and eccentricities for known planets orbiting M-dwarf
1744
+ stars (Teff < 3700). TOI-2257 b, shown with a dark marker and thick
1745
+ error bar for emphasis, is the most eccentric transiting planet with
1746
+ only one known planet in the system. Data were retrieved from the
1747
+ Exoplanet Archive (https://exoplanetarchive.ipac.caltech.
1748
+ edu/); we note that systems without reported eccentricity errors are
1749
+ shown without error bars.
1750
+
1751
+ have dynamically hotter orbits with larger eccentricities than
1752
+ those in multi-planet systems (Xie et al. 2016; Van Eylen et al.
1753
+ 2019; Masuda et al. 2020). The high eccentricity of nearly 0.5
1754
+ seen in TOI-2257 b is suggestive of the possible influence of a
1755
+ long-period giant planet rather than self-excitation, the latter of
1756
+ which Van Eylen et al. (2019) found to only explain eccentric-
1757
+ ities of up to around 0.3. An example of this mechanism was
1758
+ found for GJ 1148 b, a Saturn-mass planet orbiting an M dwarf in
1759
+ an eccentric orbit of
1760
+ 0.38 (Haghighipour et al. 2010). Posterior
1761
+ observations found that an outer giant planet was responsible for
1762
+ exciting GJ 1148 b (Trifonov et al. 2018). TOI-2257 b’s system
1763
+ parameters are consistent with the paradigm presented by Huang
1764
+ et al. (2017), in which systems that survive scattering by an outer
1765
+ giant planet show lower multiplicity and higher eccentricity.
1766
+ Future RV measurements would help to test this interpretation
1767
+ by searching for a massive planet in a more distant and likely
1768
+ non-transiting orbit.
1769
+
1770
+ ∼
1771
+
1772
+ 7.3. Prospects for radial velocity observations
1773
+
1774
+ ∼
1775
+
1776
+ 1 m s−
1777
+
1778
+ In order to constrain the mass of TOI-2257b, high-resolution
1779
+ 1 or less is needed.
1780
+ spectroscopy with a precision of
1781
+ Demangeon et al. (2021) showed in the case of the M3V-star
1782
+ 1 semi-amplitude
1783
+ L 98-59, the mass of a planet with a 2 m s−
1784
+ can be constrained with about 10% precision, taking the intrin-
1785
+ sic stellar RV jitter into account. TOI-2257b is >3 mag fainter
1786
+ than L 98-59, and thus the RV precision is reduced to about
1787
+ 1 for a 900 s exposure with a state-of-the art instru-
1788
+ 6–10 m s−
1789
+ ment such as ESPRESSO at the VLT (Pepe et al. 2021). Current
1790
+ and upcoming infrared spectrographs such as CRIRES (Dorn
1791
+ et al. 2014), SPIROU (Donati et al. 2020), or NIRPS (Wildi
1792
+ et al. 2017), will be able to obtain a higher signal-to-noise, but
1793
+ the RV precision will still not allow direct measurement of the
1794
+ reflex motion of TOI-2257. However, the newly commissioned
1795
+ MAROON-X instrument (Seifahrt et al. 2020) at the Gemini-
1796
+ 1 for
1797
+ North telescope is expected to achieve a signal of 1 m s−
1798
+ TOI-2257. Combining the instrument noise with an estimated
1799
+
1800
+ A45, page 12 of 18
1801
+
1802
+ Fig. 10. Expected semi-amplitude of a possible giant planet orbiting
1803
+ TOI-2257 for different orbits and planetary masses.
1804
+
1805
+ 1, it would be possible to constrain the
1806
+ stellar jitter of 1.5 m s−
1807
+ mass to within 6σ after 26 measurements with the red arm and
1808
+ an exposure time of 20 min.
1809
+
1810
+ Even without high-precision instruments, lower-precision
1811
+ RV follow-up might result in the detection of a more massive
1812
+ planet that is not transiting or farther out and, thus, not visible
1813
+ in the current photometric data set. We explored this possi-
1814
+ ble case since the presence of such a planet might cause the
1815
+ eccentricity of TOI-2257 b. As shown in Fig. 10, a giant planet
1816
+ with 1 MJup and orbiting in 1 au would result in an RV sig-
1817
+ 1 with an orbital period of 640 days. The
1818
+ nal of about 50 m s−
1819
+ magenta and green areas in Fig. 10 show that the detection of an
1820
+ outer massive planet will be challenging but possible for current
1821
+ high-resolution instruments. The expected long periods of such
1822
+ planets would require continuous monitoring over the course of
1823
+ months to several years with high-precision spectrographs.
1824
+
1825
+ 7.4. Prospects for atmospheric characterization
1826
+
1827
+ The long orbit of Planet b (35.9 days) makes it a rare object
1828
+ among currently known sub-Neptunes and, thus, a compelling
1829
+ target to study for its atmospheric composition and to gain
1830
+ clues about its formation and evolution. The planet’s low time-
1831
+ averaged equilibrium temperature (Méndez & Rivera-Valentín
1832
+ 2017) of Teq = 256+61
1833
+ 17 K (assuming zero albedo and full day-
1834
+ night heat redistribution) along with the star’s relative proximity
1835
+ to Earth (57.79 pc) and optical and NIR brightness (K-mag 10.67)
1836
+ strongly favors transmission spectroscopy. For example, JWST’s
1837
+ transmission spectroscopy metric (TSM) is 32.71, following
1838
+ Kempton et al. (2018).
1839
+
1840
+ −
1841
+
1842
+ Thus, TOI-2257 b could be an interesting target to study
1843
+ ocean loss on this class of temperate sub-Neptunes. Probing
1844
+ O3 abundances with red-sensitive JWST observations as well
1845
+ as O2 abundances with ground-based optical observations from
1846
+ the Extremely Large Telescopes (ELT; Luger & Barnes 2015;
1847
+ Serindag & Snellen 2019), or probing CO and O4 features
1848
+ (Schwieterman et al. 2016), could give clues in this regard.
1849
+
1850
+ ∼
1851
+
1852
+ Finally, TOI-2257 b is one of only
1853
+
1854
+ 20 currently known,
1855
+ characterizable sub-Neptunes whose equilibrium temperatures
1856
+ fall into a liquid-water-is-possible zone (Fig. 11). Of course, such
1857
+ exoplanets do not allow surface habitability, given their immense
1858
+ atmospheric pressure and heated surface. However, life in the
1859
+ clouds could be a possibility – as first proposed by Morowitz &
1860
+ Sagan (1967); Sagan & Salpeter (1976) and recently discussed
1861
+ for sub-Neptune-sized exoplanets by Seager et al. (2021) (see
1862
+
1863
+ 0.00.51.01.4Planet mass [MJup]0.10.30.60.81.1Semimajor axis [au]Semi amplitude [m s1]102030405060708090100110120 N. Schanche et al.: TOI-2257 b
1864
+
1865
+ Fig. 11. TOI-2257 b and known exoplanets in the context of their equi-
1866
+ librium temperature (x axis) and TSM (y axis). Different symbols and
1867
+ colors designate different planet size categories. TOI-2257 b lies in the
1868
+ middle of the blue shaded area, which highlights sub-Neptunes (i) that
1869
+ are amenable for atmospheric characterization (TSM > 12) and (ii) for
1870
+ which life in the clouds is possible (Teq between 200 and 320 K, follow-
1871
+ ing Seager et al. 2021). Equilibrium temperatures were estimated using
1872
+ an albedo of 0 and an emissivity of 1, following Kempton et al. (2018).
1873
+ Altering the assumptions for albedo and emissivity (e.g., assuming an
1874
+ albedo of 0.3, as for Earth and Neptune) adds lower error bars of
1875
+ 10%
1876
+ and upper error bars of
1877
+ 30% to the shown temperatures. Data were
1878
+ retrieved from the Exoplanet Archive (https://exoplanetarchive.
1879
+ ipac.caltech.edu/) on 2021 July 30. The figure has been adapted
1880
+ from Seager et al. (2021).
1881
+
1882
+ ∼
1883
+
1884
+ ∼
1885
+
1886
+ also Fig. 11). Additionally, this temperature regime could favor
1887
+ the surface habitability conditions of any potential exomoons, as
1888
+ discussed in the following section.
1889
+
1890
+ 7.5. Prospects of exomoons orbiting TOI-2257 b
1891
+
1892
+ Moons in the Solar System are invaluable puzzle pieces for
1893
+ understanding our system formation history, evolution, and even
1894
+ habitability (see, e.g., Öpik 1960; Laskar et al. 1993; Rufu
1895
+ et al. 2017; Orgel et al. 2018). Equally valuable is studying
1896
+ moons hosted in other planetary systems, that is, exomoons.
1897
+ Indeed, due to the large number of moons orbiting Solar System
1898
+ planets, the possible existence of moons orbiting exoplanets is
1899
+ widely accepted. However, despite efforts to find them (see, e.g.,
1900
+ Kipping et al. 2012, 2015; Hippke 2015; Teachey & Kipping
1901
+ 2018; Kreidberg et al. 2019), at the time of writing no detection
1902
+ has yet been confirmed.
1903
+
1904
+ One of the major interests of exomoons is their potential
1905
+ habitability. Indeed, massive rocky exomoons orbiting temper-
1906
+ ate giant planets have been suggested as places where life might
1907
+ arise and evolve over long timescales (see, e.g., Williams et al.
1908
+ 1997; Kaltenegger 2010; Heller & Barnes 2013, 2015; Heller
1909
+ et al. 2014). In particular, temperate moon–planet systems hosted
1910
+ by low-mass stars are widely discussed in the literature, provid-
1911
+ ing arguments in favor and against the existence and habitability
1912
+ of exomoons. For example, it has been suggested that such
1913
+ moons will be likely tidally locked to the planet, a situation that
1914
+ would favor a uniform distribution of irradiation (Trifonov et al.
1915
+
1916
+ Fig. 12. MSR upper limit around TOI-2257 b as a function of the plan-
1917
+ etary mass. The blue line represents the solution when considering
1918
+ TOI-2257 b as a giant planet, while the yellow line is considering it as
1919
+ a rocky planet. Each data point is the MSR computed by simulating
1920
+ 1000 exomoons orbiting around TOI-2257 b, for a total of 64 000
1921
+ scenarios tested.
1922
+
1923
+ 2020). In addition, the magnetic field generated by the planet
1924
+ may protect the moon from the typical stellar flares produced
1925
+ by M dwarfs (Heller & Zuluaga 2013). These factors would
1926
+ mitigate some typical inconveniences when discussing the hab-
1927
+ itability of a planet orbiting in the habitable zone of an M dwarf
1928
+ (Khodachenko et al. 2007; Lammer et al. 2007). On the other
1929
+ hand, gravitational perturbations by the close star and extra plan-
1930
+ ets in the system could induce eccentricities that would likely
1931
+ make any moon uninhabitable (Heller 2012). In any case, due
1932
+ to the number of currently unknown parameters, the field of
1933
+ habitable exomoons is still very speculative.
1934
+
1935
+ In this study we focus only on estimating the existence
1936
+ of rocky exomoons orbiting TOI-2257 b. To this end, we used
1937
+ dynamical considerations to compute the moon survival rate
1938
+ (MSR) following the formulation introduced by Sasaki et al.
1939
+ (2012) and the prescriptions given by Dobos et al. (2021). The
1940
+ simulation details are provided in Appendix B. The mathemati-
1941
+ cal description takes into account tidal evolution in a star–planet–
1942
+ moon system. It is important to notice that this formulation
1943
+ assumes that the planet resides in a circular orbit around the host
1944
+ star; however, this is not the case for TOI-2257 b. Hence, the
1945
+ results yielded by our simulations should be considered as upper
1946
+ limits to the actual MSR as eccentric planetary orbits reduce the
1947
+ stability of the moon–planet systems.
1948
+
1949
+ The results of the simulations are displayed in Fig. 12. In
1950
+ summary, we found that the larger the mass of the planet, the
1951
+ larger the MSR, with a maximum MSR of about 20% for ice
1952
+ and gas giant planets. On the other hand, the MSR was almost
1953
+ null for the entire sample of possible planetary masses when
1954
+ considering a rocky planet. In reality, when increasing the plan-
1955
+ etary mass, the density also increases, passing from an ice and
1956
+ gas giant to a rocky planet. This change would happen between
1957
+ 5.5 and 6.5 M
1958
+ , which, combined with the planetary radius of
1959
+ 3. The MSR
1960
+ 2.194 R
1961
+ ⊕
1962
+ 6 M
1963
+ would follow the ice and gas giant solution until
1964
+ , drop-
1965
+ ping then to the estimated MSR for rocky planets. In such a case,
1966
+ the maximum MSR given for
1967
+
1968
+ 6 M
1969
+ ⊕
1970
+ We conclude that TOI-2257 b is likely a single planet with a
1971
+ low probability of having any orbiting moons. This result is in
1972
+ line with those found by Sasaki et al. (2012) and Dobos et al.
1973
+ (2021), who established that it is very challenging for planets
1974
+ orbiting low-mass stars with a semimajor axis
1975
+ 0.2 au to harbor
1976
+ moons in stable orbits.
1977
+
1978
+ , yields a planetary density of 2.9–3.5 g cm−
1979
+
1980
+ is about 13%.
1981
+
1982
+ ∼
1983
+
1984
+ ≤
1985
+
1986
+ ∼
1987
+
1988
+ ⊕
1989
+
1990
+ ⊕
1991
+
1992
+ A45, page 13 of 18
1993
+
1994
+ 20050010002000Planet Equilibrium Temperature (Kelvin)100101102103Transmission Spectroscopy MetricTOI-2557 b< 1.5 Re< 1.8 Re< 2.75 Re< 4 Re> 4 Re34567891011TOI-2257b Mass (M)0510152025Moon Survival Rate (%)Ice/Gas giantRocky A&A 657, A45 (2022)
1995
+
1996
+ 8. Conclusion
1997
+
1998
+ This work presents the discovery and characterization of TOI-
1999
+ 2257 b, a sub-Neptune in a relatively long-period orbit around an
2000
+ M3V star. The preliminary characterization is based on photom-
2001
+ etry from TESS and several ground-based facilities, with spectral
2002
+ analysis and high-resolution imaging supporting the validation.
2003
+ TOI-2257 b occupies a sparsely populated region of parameter
2004
+ space. Currently, only two other transiting exoplanets around
2005
+ M dwarfs with periods greater than that of TOI-2257 b are
2006
+ known (Kepler-1652 b and TOI-700 d; Torres et al. 2017; Gilbert
2007
+ et al. 2020; Rodriguez et al. 2020). Furthermore, this single-
2008
+ planet system is the most eccentric planet transiting an M dwarf
2009
+ discovered to date, providing an opportunity to test possible
2010
+ formation scenarios.
2011
+
2012
+ While the expected driven reflex motion, on the order of a
2013
+ 1, makes current observations challenging, recent and
2014
+ few m s−
2015
+ upcoming instrumentation at observatories with larger collect-
2016
+ ing areas, such as MAROON-X at the Gemini Observatory and
2017
+ the HIRES spectrograph at the ELT (Marconi et al. 2021), will
2018
+ make it possible to obtain spectra for TOI-2257 with more than
2019
+ 20 times higher S/N than with ESPRESSO. Thus, these large
2020
+ telescopes will enable us to derive the mass of TOI-2257 b to
2021
+ within 10% and even allow us to search for possible other plan-
2022
+ ets within the system. Regardless, this planet is one of only a
2023
+ small number of sub-Neptunes for which liquid water is a pos-
2024
+ sibility. The expected TSM of the object makes it amenable to
2025
+ more detailed atmospheric characterization from JWST.
2026
+
2027
+ Acknowledgements. N.S., R.W. and B.-O.D. acknowledge support from the Swiss
2028
+ National Science Foundation (PP00P2-163967 and PP00P2-190080). M.N.G.
2029
+ acknowledges support from MIT’s Kavli Institute as a Juan Carlos Torres Fel-
2030
+ low and from the European Space Agency (ESA) as an ESA Research Fellow.
2031
+ A.A.B., B.S.S.and I.A.S. acknowledge the support of the Ministry of Science and
2032
+ Higher Education of the Russian Federation under the grant 075-15-2020-780
2033
+ (N13.1902.21.0039). L.D. is an F.R.S.-FNRS Postdoctoral Researcher. B.V.R.
2034
+ thanks the Heising-Simons Foundation for support. This publication benefits
2035
+ from the support of the French Community of Belgium in the context of the
2036
+ FRIA Doctoral Grant awarded to M.T. and E.J. acknowledges DGAPA for his
2037
+ postdoctoral fellowship. Y.G.M.C. acknowledges support from UNAM-DGAPA
2038
+ PAPIIT BG-101321. D.D. acknowledges support from the TESS Guest Investi-
2039
+ gator Program grant 80NSSC19K1727 and NASA Exoplanet Research Program
2040
+ grant 18-2XRP18_2-0136. We acknowledge support from the Centre for Space
2041
+ and Habitability (CSH) of the University of Bern. Part of this work received sup-
2042
+ port from the National Centre for Competence in Research PlanetS, supported
2043
+ by the Swiss National Science Foundation (SNSF). Funding for the TESS mis-
2044
+ sion is provided by NASA’s Science Mission Directorate. We acknowledge the
2045
+ use of public TESS data from pipelines at the TESS Science Office and at
2046
+ the TESS Science Processing Operations Center. This research has made use of
2047
+ the Exoplanet Follow-up Observation Program website, which is operated by the
2048
+ California Institute of Technology, under contract with the National Aeronautics
2049
+ and Space Administration under the Exoplanet Exploration Program. Resources
2050
+ supporting this work were provided by the NASA High-End Computing (HEC)
2051
+ Program through the NASA Advanced Supercomputing (NAS) Division at Ames
2052
+ Research Center for the production of the SPOC data products. This paper
2053
+ includes data collected by the TESS mission that are publicly available from
2054
+ the Mikulski Archive for Space Telescopes (MAST). This work is based upon
2055
+ observations carried out at the Observatorio Astronómico Nacional on the Sierra
2056
+ de San Pedro Mártir (OAN-SPM), Baja California, México. We warmly thank
2057
+ the entire technical staff of the Observatorio Astronómico Nacional at San
2058
+ Pedro Mártir in México for their unfailing support to SAINT-EX operations,
2059
+ namely: E. Cadena, T. Calvario, E. Colorado, F. Díaz, A. Franco, B. García,
2060
+ C. Guerrero, G. Guisa, F. Guillen, A. Landa, L. Figueroa, B. Hernández, J. Her-
2061
+ rera, E. López, E. Lugo, B. Martínez, G. Melgoza, F. Montalvo, J.M. Nuñez,
2062
+ J.L. Ochoa, I. Plauchu, F. Quiroz, H. Riesgo, H. Serrano, T. Verdugo, I. Zavala.
2063
+ The research leading to these results has received funding from the Euro-
2064
+ pean Research Council (ERC) under the FP/2007–2013 ERC grant agreement
2065
+ n◦ 336480, and under the European Union’s Horizon 2020 research and
2066
+ innovation programme (grants agreements n◦ 679030 and 803193/BEBOP);
2067
+ from an Actions de Recherche Concertée (ARC) grant, financed by the
2068
+ Wallonia–Brussels Federation, from the Balzan Prize Foundation, from the BEL-
2069
+ SPO/BRAIN2.0 research program (PORTAL project), from the Science and
2070
+
2071
+ A45, page 14 of 18
2072
+
2073
+ Technology Facilities Council (STFC; grant n◦ ST/S00193X/1), and from F.R.S-
2074
+ FNRS (Research Project ID T010920F). This work was also partially supported
2075
+ by a grant from the Simons Foundation (PI: Queloz, grant number 327127),
2076
+ as well as by the MERAC foundation (PI: Triaud). PI: Gillon is F.R.S.-FNRS
2077
+ Senior Research Associate. TRAPPIST is funded by the Belgian Fund for Sci-
2078
+ entific Research (Fond National de la Recherche Scientifique, FNRS) under the
2079
+ grant PDR T.0120.21, with the participation of the Swiss National Science Fun-
2080
+ dation (SNF). M.G. and E.J. are F.R.S.-FNRS Senior Research Associate. This
2081
+ work makes use of observations from the LCOGT network. Part of the LCOGT
2082
+ telescope time was granted by NOIRLab through the Mid-Scale Innovations Pro-
2083
+ gram (MSIP). M.S.I.P. is funded by NSF. Some of the observations in the paper
2084
+ made use of the High-Resolution Imaging instrument(s) ‘Alopeke (and/or Zorro).
2085
+ ‘Alopeke (and/or Zorro) was funded by the NASA Exoplanet Exploration Pro-
2086
+ gram and built at the NASA Ames Research Center by Steve B. Howell, Nic
2087
+ Scott, Elliott P. Horch, and Emmett Quigley. Data were reduced using a software
2088
+ pipeline originally written by Elliott Horch and Mark Everett. ‘Alopeke (and/or
2089
+ Zorro) was mounted on the Gemini North (and/or South) telescope of the inter-
2090
+ national Gemini Observatory, a program of NSF’s OIR Lab, which is managed
2091
+ by the Association of Universities for Research in Astronomy (AURA) under a
2092
+ cooperative agreement with the National Science Foundation, on behalf of the
2093
+ Gemini partnership: the National Science Foundation (United States), National
2094
+ Research Council (Canada), Agencia Nacional de Investigación y Desarrollo
2095
+ (Chile), Ministerio de Ciencia, Tecnología e Innovación (Argentina), Ministério
2096
+ da Ciência, Tecnologia, Inovações e Comunicações (Brazil), and Korea Astron-
2097
+ omy and Space Science Institute (Republic of Korea). This research has made use
2098
+ of the NASA Exoplanet Archive, which is operated by the California Institute of
2099
+ Technology, under contract with the National Aeronautics and Space Admin-
2100
+ istration under the Exoplanet Exploration Program. This research made use of
2101
+ exoplanet (Foreman-Mackey et al. 2021a,b) and its dependencies (Agol et al.
2102
+ 2020; Kumar et al. 2019; Astropy Collaboration 2013, 2018; Kipping 2013; Luger
2103
+ et al. 2019; Salvatier et al. 2016; Theano Development Team 2016). Additional
2104
+ use of software packages AstroImageJ (Collins et al. 2017) and TAPIR (Jensen
2105
+ 2013).
2106
+
2107
+ References
2108
+
2109
+ Agol, E., Luger, R., & Foreman-Mackey, D. 2020, AJ, 159, 123
2110
+ Aller, A., Lillo-Box, J., Jones, D., Miranda, L. F., & Barceló Forteza, S. 2020,
2111
+
2112
+ A&A, 635, A128
2113
+
2114
+ Alonso-Floriano, F. J., Morales, J. C., Caballero, J. A., et al. 2015, A&A, 577,
2115
+
2116
+ A128
2117
+
2118
+ Astropy Collaboration (Robitaille, T. P., et al.) 2013, A&A, 558, A33
2119
+ Astropy Collaboration (Price-Whelan, A. M., et al.) 2018, AJ, 156, 123
2120
+ Badenas-Agusti, M., Günther, M. N., Daylan, T., et al. 2020, AJ, 160, 113
2121
+ Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R.
2122
+
2123
+ 2018, AJ, 156, 58
2124
+
2125
+ Baraffe, I., Homeier, D., Allard, F., & Chabrier, G. 2015, A&A, 577, A42
2126
+ Barkaoui, K., Burdanov, A., Hellier, C., et al. 2019, AJ, 157, 43
2127
+ Bonfils, X., Delfosse, X., Udry, S., et al. 2013, A&A, 549, A109
2128
+ Borucki, W. J., Koch, D., Basri, G., et al. 2010, Science, 327, 977
2129
+ Boss, A. P. 2001, ApJ, 551, L167
2130
+ Brown, T. M., Baliber, N., Bianco, F. B., et al. 2013, PASP, 125, 1031
2131
+ Buchner, J., Georgakakis, A., Nandra, K., et al. 2014, A&A, 564, A125
2132
+ Casagrande, L., Flynn, C., & Bessell, M. 2008, MNRAS, 389, 585
2133
+ Chen, J., & Kipping, D. 2017, ApJ, 834, 17
2134
+ Choi, J., Dotter, A., Conroy, C., et al. 2016, ApJ, 823, 102
2135
+ Cloutier, R., Rodriguez, J. E., Irwin, J., et al. 2020, AJ, 160, 22
2136
+ Collins, K. A., Kielkopf, J. F., Stassun, K. G., & Hessman, F. V. 2017, AJ, 153,
2137
+
2138
+ 77
2139
+
2140
+ Covey, K. R., Ivezi´c, Ž., Schlegel, D., et al. 2007, AJ, 134, 2398
2141
+ Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, 2MASS All Sky Catalog
2142
+
2143
+ of point sources
2144
+
2145
+ Cutri, R. M., Wright, E. L., Conrow, T., et al. 2021, VizieR Online Data Catalog:
2146
+
2147
+ II/328
2148
+
2149
+ Daylan, T., Pinglé, K., Wright, J., et al. 2021, AJ, 161, 85
2150
+ Delrez, L., Gillon, M., Queloz, D., et al. 2018, SPIE, 10700, 446
2151
+ Demangeon, O. D. S., Zapatero Osorio, M. R., Alibert, Y., et al. 2021, A&A,
2152
+
2153
+ 653, A41
2154
+
2155
+ Demory, B. O., Pozuelos, F. J., Gómez Maqueo Chew, Y., et al. 2020, A&A,
2156
+
2157
+ 642, A49
2158
+
2159
+ Dobos, V., Charnoz, S., Pál, A., Roque-Bernard, A., & Szabó, G. M. 2021, PASP,
2160
+
2161
+ 133, 094401
2162
+
2163
+ Donati, J. F., Kouach, D., Moutou, C., et al. 2020, MNRAS, 498, 5684
2164
+ Dorn, R. J., Anglada-Escude, G., Baade, D., et al. 2014, The Messenger, 156, 7
2165
+ Dotter, A. 2016, ApJS, 222, 8
2166
+ Eisner, N. L., Barragán, O., Aigrain, S., et al. 2020, MNRAS, 494, 750
2167
+ Espinoza, N., Kossakowski, D., & Brahm, R. 2019, MNRAS, 490, 2262
2168
+
2169
+ N. Schanche et al.: TOI-2257 b
2170
+
2171
+ Feroz, F., Hobson, M. P., & Bridges, M. 2009, MNRAS, 398, 1601
2172
+ Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125,
2173
+
2174
+ 306
2175
+
2176
+ Foreman-Mackey, D., Luger, R., Agol, E., et al. 2021a, J. Open Source Softw., 6,
2177
+
2178
+ 3285
2179
+
2180
+ Murray, C. A., Delrez, L., Pedersen, P. P., et al. 2020, MNRAS, 495, 2446
2181
+ Niraula, P., Wit, J. d., Rackham, B. V., et al. 2020, AJ, 160, 172
2182
+ Nutzman, P., & Charbonneau, D. 2008, PASP, 120, 317
2183
+ Öpik, E. J. 1960, MNRAS, 120, 404
2184
+ Orgel, C., Michael, G., Fassett, C. I., et al. 2018, J. Geophys. Res. Planets, 123,
2185
+
2186
+ Foreman-Mackey, D., Savel, A., Luger, R., et al. 2021b, exoplanet-dev/exoplanet
2187
+
2188
+ 748
2189
+
2190
+ v0.5.1
2191
+
2192
+ Gaia Collaboration (Prusti, T., et al.) 2016, A&A, 595, A1
2193
+ Gaia Collaboration (Brown, A. G. A., et al.) 2018, A&A, 616, A1
2194
+ Gaia Collaboration (Brown, A. G. A., et al.) 2021, A&A, 649, A1
2195
+ Garcia, L. J., Timmermans, M., Pozuelos, F. J., et al. 2021, MNRAS, 509, 4817
2196
+ Giacalone, S., Dressing, C. D., Jensen, E. L. N., et al. 2021, AJ, 161, 24
2197
+ Gilbert, E. A., Barclay, T., Schlieder, J. E., et al. 2020, AJ, 160, 116
2198
+ Gillon, M., Jehin, E., Magain, P., et al. 2011, EPJ Web Conf., 11, 06002
2199
+ Goodman, J., & Weare, J. 2010, Commun. Appl. Math. Comput. Sci., 5, 65
2200
+ Guerrero, N. M., Seager, S., Huang, C. X., et al. 2021, ApJS, 254, 39
2201
+ Haghighipour, N., Vogt, S. S., Butler, R. P., et al. 2010, ApJ, 715, 271
2202
+ Hauschildt, P. H., Allard, F., & Baron, E. 1999, ApJ, 512, 377
2203
+ Hedges, C., Hughes, A., Zhou, G., et al. 2021, AJ, 162, 54
2204
+ Heller, R. 2012, A&A, 545, L8
2205
+ Heller, R., & Barnes, R. 2013, Astrobiology, 13, 18
2206
+ Heller, R., & Barnes, R. 2015, Int. J. Astrobiol., 14, 335
2207
+ Heller, R., & Zuluaga, J. I. 2013, ApJ, 776, L33
2208
+ Heller, R., Williams, D., Kipping, D., et al. 2014, Astrobiology, 14, 798
2209
+ Henden, A. A., Levine, S., Terrell, D., & Welch, D. L. 2015, AAS Meeting Abs.,
2210
+
2211
+ Parviainen, H. 2015, MNRAS, 450, 3233
2212
+ Pepe, F., Cristiani, S., Rebolo, R., et al. 2021, A&A, 645, A96
2213
+ Pozuelos, F. J., Suárez, J. C., de Elía, G. C., et al. 2020, A&A, 641, A23
2214
+ Quinn, S. N., Becker, J. C., Rodriguez, J. E., et al. 2019, AJ, 158, 177
2215
+ Rains, A. D., Žerjal, M., Ireland, M. J., et al. 2021, MNRAS, 504, 5788
2216
+ Reiners, A., Zechmeister, M., Caballero, J. A., et al. 2018, A&A, 612, A49
2217
+ Richer, H. B., Anderson, J., Brewer, J., et al. 2006, Science, 313, 936
2218
+ Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2015, J. Astron. Teles. Instrum.
2219
+
2220
+ Syst., 1, 014003
2221
+
2222
+ Rodriguez, J. E., Vanderburg, A., Zieba, S., et al. 2020, AJ, 160, 117
2223
+ Rufu, R., Aharonson, O., & Perets, H. B. 2017, Nat. Geosci., 10, 89
2224
+ Safonov, B. S., Lysenko, P. A., & Dodin, A. V. 2017, Astron. Lett., 43, 344
2225
+ Sagan, C., & Salpeter, E. E. 1976, ApJS, 32, 737
2226
+ Salvatier, J., Wiecki, T. V., & Fonnesbeck, C. 2016, PeerJ Comput. Sci., 2, e55
2227
+ Sasaki, T., Barnes, J. W., & O’Brien, D. P. 2012, ApJ, 754, 51
2228
+ Schlaufman, K. C., & Laughlin, G. 2010, A&A, 519, A105
2229
+ Schwieterman, E. W., Meadows, V. S., Domagal-Goldman, S. D., et al. 2016,
2230
+
2231
+ ApJ, 819, L13
2232
+
2233
+ Scott, N. J., Howell, S. B., Gnilka, C. L., et al. 2021, Front. Astron. Space Sci.,
2234
+
2235
+ 225, 336.16
2236
+
2237
+ 8, 138
2238
+
2239
+ Hippke, M. 2015, ApJ, 806, 51
2240
+ Hippke, M., & Heller, R. 2019, A&A, 623, A39
2241
+ Hippke, M., David, T. J., Mulders, G. D., & Heller, R. 2019, AJ, 158, 143
2242
+ Howell, S. B., Everett, M. E., Sherry, W., Horch, E., & Ciardi, D. R. 2011, AJ,
2243
+
2244
+ 142, 19
2245
+
2246
+ Howell, S. B., Sobeck, C., Haas, M., et al. 2014, PASP, 126, 398
2247
+ Howell, S. B., Everett, M. E., Horch, E. P., et al. 2016, ApJ, 829, L2
2248
+ Huang, C. X., Petrovich, C., & Deibert, E. 2017, AJ, 153, 210
2249
+ Jehin, E., Gillon, M., Queloz, D., et al. 2011, The Messenger, 145, 2
2250
+ Jenkins, J. M. 2002, ApJ, 575, 493
2251
+ Jenkins, J. M., Chandrasekaran, H., McCauliff, S. D., et al. 2010, SPIE Conf.
2252
+
2253
+ Ser., 7740, 77400D
2254
+
2255
+ Jenkins, J. M., Twicken, J. D., McCauliff, S., et al. 2016, Proc. SPIE, 9913,
2256
+
2257
+ 99133E
2258
+
2259
+ Jenkins, J. M., Tenenbaum, P., Seader, S., et al. 2020, Kepler Data Processing
2260
+
2261
+ Handbook: Transiting Planet Search (USA: Kepler Science Document)
2262
+ Jensen, E. 2013, Astrophysics Source Code Library [record ascl:1306.007]
2263
+ Kaltenegger, L. 2010, ApJ, 712, L125
2264
+ Kass, R. E., & Raftery, A. E. 1995, J. Am. Stat. Assoc., 90, 773
2265
+ Kempton, E. M. R., Bean, J. L., Louie, D. R., et al. 2018, PASP, 130, 114401
2266
+ Kesseli, A. Y., West, A. A., Veyette, M., et al. 2017, ApJS, 230, 16
2267
+ Khodachenko, M. L., Ribas, I., Lammer, H., et al. 2007, Astrobiology, 7, 167
2268
+ Kipping, D. M. 2013, MNRAS, 435, 2152
2269
+ Kipping, D. M., Bakos, G. Á., Buchhave, L., Nesvorný, D., & Schmitt, A. 2012,
2270
+
2271
+ ApJ, 750, 115
2272
+
2273
+ Kipping, D. M., Schmitt, A. R., Huang, X., et al. 2015, ApJ, 813, 14
2274
+ Kreidberg, L. 2015, PASP, 127, 1161
2275
+ Kreidberg, L., Luger, R., & Bedell, M. 2019, ApJ, 877, L15
2276
+ Kumar, R., Carroll, C., Hartikainen, A., & Martin, O. A. 2019, J. Open Source
2277
+
2278
+ Softw.
2279
+
2280
+ Lammer, H., Lichtenegger, H. I. M., Kulikov, Y. N., et al. 2007, Astrobiology, 7,
2281
+
2282
+ 185
2283
+
2284
+ Laskar, J., Joutel, F., & Robutel, P. 1993, Nature, 361, 615
2285
+ Lépine, S., Rich, R. M., & Shara, M. M. 2007, ApJ, 669, 1235
2286
+ Li, J., Tenenbaum, P., Twicken, J. D., et al. 2019, PASP, 131, 024506
2287
+ Luger, R., & Barnes, R. 2015, Astrobiology, 15, 119
2288
+ Luger, R., Agol, E., Foreman-Mackey, D., et al. 2019, AJ, 157, 64
2289
+ Mandel, K., & Agol, E. 2002, ApJ, 580, L171
2290
+ Mann, A. W., Brewer, J. M., Gaidos, E., Lépine, S., & Hilton, E. J. 2013, AJ,
2291
+
2292
+ 145, 52
2293
+
2294
+ Mann, A. W., Dupuy, T., Kraus, A. L., et al. 2019, ApJ, 871, 63
2295
+ Marconi, A., Abreu, M., Adibekyan, V., et al. 2021, The Messenger, 182, 27
2296
+ Masuda, K., Winn, J. N., & Kawahara, H. 2020, AJ, 159, 38
2297
+ Massey, P., & Gronwall, C. 1990, ApJ, 358, 344
2298
+ McCully, C., Volgenau, N. H., Harbeck, D.-R., et al. 2018, Proc. SPIE Conf. Ser.,
2299
+
2300
+ 10707, 107070K
2301
+
2302
+ Méndez, A., & Rivera-Valentín, E. G. 2017, ApJ, 837, L1
2303
+ Morowitz, H., & Sagan, C. 1967, Nature, 215, 1259
2304
+ Morton, T. D.
2305
+
2306
+ 2015, Astrophysics Source Code Library
2307
+
2308
+ [record
2309
+
2310
+ ascl:1503.010]
2311
+
2312
+ Muirhead, P. S., Dressing, C. D., Mann, A. W., et al. 2018, VizieR Online Data
2313
+
2314
+ Catalog: J/AJ/155/180
2315
+
2316
+ Seager, S., Petkowski, J. J., Günther, M. N., et al. 2021, Universe, 7, 172
2317
+ Sebastian, D., Gillon, M., Ducrot, E., et al. 2021, A&A, 645, A100
2318
+ Seifahrt, A., Bean, J. L., Stürmer, J., et al. 2020, SPIE, 11447, 305
2319
+ Serindag, D. B., & Snellen, I. A. G. 2019, ApJ, 871, L7
2320
+ Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163
2321
+ Slesnick, C. L., Carpenter, J. M., & Hillenbrand, L. A. 2006, AJ, 131, 3016
2322
+ Smith, J. C., Stumpe, M. C., Van Cleve, J. E., et al. 2012, PASP, 124, 1000
2323
+ Speagle, J. S. 2020, MNRAS, 493, 3132
2324
+ Stassun, K. G., & Torres, G. 2016, AJ, 152, 180
2325
+ Stassun, K. G., & Torres, G. 2021, ApJ, 907, L33
2326
+ Stassun, K. G., Collins, K. A., & Gaudi, B. S. 2017, AJ, 153, 136
2327
+ Stassun, K. G., Corsaro, E., Pepper, J. A., & Gaudi, B. S. 2018a, AJ, 155, 22
2328
+ Stassun, K. G., Oelkers, R. J., Pepper, J., et al. 2018b, AJ, 156, 102
2329
+ Stassun, K. G., Oelkers, R. J., Paegert, M., et al. 2019, AJ, 158, 138
2330
+ Stumpe, M. C., Smith, J. C., Van Cleve, J. E., et al. 2012, PASP, 124, 985
2331
+ Stumpe, M. C., Smith, J. C., Catanzarite, J. H., et al. 2014, PASP, 126, 100
2332
+ Teachey, A., & Kipping, D. M. 2018, Sci. Adv., 4, eaav1784
2333
+ Theano Development Team 2016, ArXiv e-prints [arXiv:1605.02688]
2334
+ Torres, G., Kane, S. R., Rowe, J. F., et al. 2017, AJ, 154, 264
2335
+ Trifonov, T., Kürster, M., Zechmeister, M., et al. 2018, A&A, 609, A117
2336
+ Trifonov, T., Lee, M. H., Kürster, M., et al. 2020, A&A, 638, A16
2337
+ Twicken, J. D., Catanzarite, J. H., Clarke, B. D., et al. 2018, PASP, 130, 064502
2338
+ Van Eylen, V., Albrecht, S., Huang, X., et al. 2019, AJ, 157, 61
2339
+ Wells, R. D., Rackham, B. V., Schanche, N., et al. 2021, A&A, 653, A97
2340
+ Wildi, F., Blind, N., Reshetov, V., et al. 2017, SPIE Conf. Ser., 10400, 1040018
2341
+ Williams, D. M., Kasting, J. F., & Wade, R. A. 1997, Nature, 385, 234
2342
+ Woolf, V. M., & Wallerstein, G. 2006, PASP, 118, 218
2343
+ Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868
2344
+ Xie, J.-W., Dong, S., Zhu, Z., et al. 2016, Proc. Natl. Acad. Sci., 113, 11431
2345
+ Zeng, L., Sasselov, D. D., & Jacobsen, S. B. 2016, ApJ, 819, 127
2346
+
2347
+ 1 Center
2348
+
2349
+ for Space
2350
+
2351
+ and Habitability, University
2352
+
2353
+ of Bern,
2354
+
2355
+ Gesellschaftsstrasse 6, 3012 Bern, Switzerland
2356
+ e-mail: nicole.schanche@unibe.ch
2357
+
2358
+ 2 Astrobiology Research Unit, Université de Liège, Allée du 6 Août
2359
+
2360
+ 19C, 4000 Liège, Belgium
2361
+
2362
+ 3 Space Sciences, Technologies and Astrophysics Research (STAR)
2363
+ Institute, Université de Liège, Allée du 6 Août 19C, 4000 Liège,
2364
+ Belgium
2365
+
2366
+ 4 Department of Physics, and Kavli Institute for Astrophysics and
2367
+ Space Research, Massachusetts Institute of Technology (MIT),
2368
+ Cambridge, MA 02139, USA
2369
+
2370
+ 5 European Space Agency (ESA), European Space Research and
2371
+ Technology Centre (ESTEC), Keplerlaan 1, 2201 AZ Noordwijk,
2372
+ The Netherlands
2373
+
2374
+ 6 Center for Astrophysics and Space Science, University of California
2375
+
2376
+ San Diego, La Jolla, CA 92093, USA
2377
+
2378
+ A45, page 15 of 18
2379
+
2380
+ 7 Instituto de Astrofísica de Canarias (IAC), Calle Vía Láctea s/n,
2381
+
2382
+ 20 Caltech/IPAC, 1200 E. California Boulevard, Pasadena, CA 91125,
2383
+
2384
+ A&A 657, A45 (2022)
2385
+
2386
+ USA
2387
+
2388
+ 21 Dpto. Física Teórica y del Cosmos, Universidad de Granada, 18071
2389
+
2390
+ Granada, Spain
2391
+
2392
+ 22 Department of Physics and Astronomy, University of New Mexico,
2393
+
2394
+ 1919 Lomas Blvd NE, Albuquerque, NM 87131, USA
2395
+
2396
+ 23 NASA Exoplanet Science Institute, Caltech/IPAC, Mail Code 100-
2397
+
2398
+ 22, 1200 E. California Blvd., Pasadena, CA 91125, USA
2399
+
2400
+ 24 NASA Ames Research Center, Moffett Field, CA 94035, USA
2401
+ 25 Universidad Nacional Autónoma de México,
2402
+ Astronomía, AP 106, Ensenada 22800, BC, Mexico
2403
+
2404
+ Instituto de
2405
+
2406
+ 26 Department of Astronomy, University of Florida, Gainesville, FL,
2407
+
2408
+ 32611, USA
2409
+
2410
+ 27 Department of Physics, University of Warwick, Gibbet Hill Road,
2411
+
2412
+ Coventry CV4 7AL, UK
2413
+
2414
+ 28 Dept. of Physics & Astronomy, Swarthmore College, Swarthmore
2415
+
2416
+ PA 19081, USA
2417
+
2418
+ 29 Department
2419
+
2420
+ of Astronomy, Wellesley College, Wellesley,
2421
+
2422
+ MA 02481, USA
2423
+
2424
+ 30 Cavendish Laboratory, JJ Thomson Avenue, Cambridge, CB3 0HE,
2425
+
2426
+ UK
2427
+
2428
+ 31 NASA Goddard Space Flight Center, 8800 Greenbelt Road,
2429
+
2430
+ Greenbelt, MD 20771, USA
2431
+
2432
+ 32 Department
2433
+
2434
+ of Aeronautics
2435
+
2436
+ and Astronautics, MIT,
2437
+
2438
+ 77
2439
+
2440
+ Massachusetts Avenue, Cambridge, MA 02139, USA
2441
+
2442
+ 33 Department of Astrophysical Sciences, Princeton University,
2443
+
2444
+ 4 Ivy Lane, Princeton, NJ 08544, USA
2445
+
2446
+ 38200 La Laguna, Tenerife, Spain
2447
+
2448
+ 8 Universidad Nacional Autónoma de México,
2449
+ Astronomía, AP 70-264, CDMX 04510, Mexico
2450
+
2451
+ Instituto de
2452
+
2453
+ 9 Universidad Nacional de Córdoba - Observatorio Astronómico de
2454
+
2455
+ Córdoba, Laprida 854, X5000BGR, Córdoba, Argentina
2456
+
2457
+ 10 Consejo Nacional de Investigaciones Científicas y Técnicas (CON-
2458
+
2459
+ ICET), Argentina
2460
+
2461
+ 11 Department of Earth, Atmospheric and Planetary Science,
2462
+ Massachusetts Institute of Technology, 77 Massachusetts Avenue,
2463
+ Cambridge, MA 02139, USA
2464
+
2465
+ 12 School of Physics & Astronomy, University of Birmingham,
2466
+
2467
+ Edgbaston, Birmimgham B15 2TT, UK
2468
+
2469
+ 13 Department of Physics & Astronomy, Vanderbilt University,
2470
+
2471
+ 6301 Stevenson Center Ln., Nashville, TN 37235, USA
2472
+
2473
+ 14 Jet Propulsion Laboratory, California Institute of Technology,
2474
+ 4800 Oak Grove Drive, MS 169-224, Pasadena, CA 91109, USA
2475
+ 15 Oukaimeden Observatory, High Energy Physics and Astrophysics
2476
+
2477
+ Laboratory, Cadi Ayyad University, Marrakech, Morocco
2478
+
2479
+ 16 Sternberg Astronomical Institute, Moscow State University (SAI
2480
+
2481
+ MSU), Universitetskii pr. 13, Moscow 119991, Russia
2482
+
2483
+ 17 Physikalisches Institut, University of Bern, Gesellsschaftstrasse 6,
2484
+
2485
+ 3012 Bern, Switzerland
2486
+
2487
+ 18 Center for Astrophysics | Harvard & Smithsonian, 60 Garden Street,
2488
+
2489
+ Cambridge, MA, 02138, USA
2490
+
2491
+ 19 Observatoire de l’Université de Genève, Chemin des Maillettes 51,
2492
+
2493
+ 1290 Versoix, Switzerland
2494
+
2495
+ A45, page 16 of 18
2496
+
2497
+ Appendix A: Posterior distribution of transit parameters
2498
+
2499
+ N. Schanche et al.: TOI-2257 b
2500
+
2501
+ Fig. A.1. MCMC posterior distributions for the fit transit parameters. From left-to-right: Transit epoch (BJD
2502
+ parameter, planet-to-star radius ratio, stellar radius (R
2503
+
2504
+ ), √e cos ω, and √e sin ω.
2505
+
2506
+ ), stellar mass (M
2507
+ ∗
2508
+
2509
+ ∗
2510
+
2511
+ 2,450,000), orbital period (d), impact
2512
+
2513
+ −
2514
+
2515
+ A45, page 17 of 18
2516
+
2517
+ 0.00900.00920.00940.0096Per+3.518e10.000.250.500.75b0.0600.0630.0660.069RpRs0.240.280.320.36R*(R)0.240.280.320.360.40M*(M)0.40.00.4ecos0.07500.07750.08000.0825T0+9.0079e30.750.600.450.30esin0.00900.00920.00940.0096Per+3.518e10.000.250.500.75b0.0600.0630.0660.069RpRs0.240.280.320.36R*(R)0.240.280.320.360.40M*(M)0.40.00.4ecos0.750.600.450.30esin A&A 657, A45 (2022)
2518
+
2519
+ Appendix B: Exomoon simulations
2520
+
2521
+ The main parameters that describe the stability of a moon are
2522
+ the planetary mass, radius, and semimajor axis, the quality factor
2523
+ that describes the dissipation energy Qp, the rotational state, and
2524
+ the potential Love number k2p.
2525
+
2526
+ ⊕
2527
+
2528
+ in steps of 0.25 M
2529
+ ⊕
2530
+
2531
+ In our case, only a few of these parameters are known:
2532
+ the planetary radius, Rp=2.194 R
2533
+ , and semimajor axis,
2534
+ ap=0.145 au. The other parameters are unknown. Hence, we ex-
2535
+ plored them by taking into account the considerations described
2536
+ in detail by Dobos et al. (2021): (i) We explored the planetary
2537
+ mass, Mp, considering its uncertainty as given in Table 3, that is,
2538
+ ranging from 3 to 10.75 M
2539
+ . (ii) The qual-
2540
+ ⊕
2541
+ ity factor highly depends on the physical characteristics of the
2542
+ planet. It was established that for Rp < 2.0 R
2543
+ (rocky planets),
2544
+ 10<Qp<500, and for Rp (cid:62) 2.0 R
2545
+ ⊕
2546
+ (ice and gas giant planets)
2547
+ ⊕
2548
+ with orbital periods larger than 10 d, 103<Qp<106, with the most
2549
+ 104. Our estimation of the
2550
+ probable value defined as Qp = 3
2551
+ planetary radius is Rp=2.194+0.113
2552
+ , which is just within the
2553
+ limit between the given definition of rocky and ice and gas giant
2554
+ planets. To avoid as many biases as possible in our study, we ran
2555
+ two suites of simulations: one considering a rocky planet with
2556
+ 102 and the other considering an ice and gas giant
2557
+ Qp = 2.5
2558
+ 104. (iii) The rotational spin of the planet was
2559
+ with Qp = 3
2560
+ randomly explored between 10 h and 5 days. (iv) Finally, the
2561
+ potential Love number was assumed to be 0.299 for rocky planets
2562
+ and 0.5 for giant planets.
2563
+
2564
+ ×
2565
+ 0.111 R
2566
+ ⊕
2567
+ −
2568
+
2569
+ ×
2570
+
2571
+ ×
2572
+
2573
+ For each scenario, we generated a sample of 1000 moons
2574
+ whose stabilities were measured over a timescale of 1 Gyr. The
2575
+ parameters of moons we explored were: (i) the moon’s mass,
2576
+ which was sampled randomly over the range (0.1–0.01)
2577
+ Mp fol-
2578
+ lowing a uniform distribution; (ii) the moon’s density, which was
2579
+ explored following a Gaussian distribution with a mean value of
2580
+ 3; and (iii) the moon’s semima-
2581
+ ρm=3 g cm−
2582
+ jor axis, am, which was chosen randomly from a uniform distri-
2583
+ bution between two times the Roche limit (Rlimit) and the critical
2584
+ distance (Cd).
2585
+
2586
+ 3 and σ = 1/3 g cm−
2587
+
2588
+ ×
2589
+
2590
+ In each case, we monitored the dynamical evolution of the
2591
+ moon, prematurely finishing the integration if the moon collided
2592
+ Rlimit
2593
+ with the planet or escaped from the system, that is, am ≤
2594
+ Cd, respectively. Then, the MSR was defined as the ratio
2595
+ or am ≥
2596
+ of surviving moons (i.e., those that stayed in orbit around the
2597
+ planet until the end of the integration) and the number of tested
2598
+ configurations. In total, we explored 64,000 different scenarios.
2599
+ The results are explored further in the main text.
2600
+
2601
+ A45, page 18 of 18
2602
+
2603
+
documents/txt/toi2406.txt ADDED
@@ -0,0 +1,2153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ A&A 653, A97 (2021)
2
+ https://doi.org/10.1051/0004-6361/202141277
3
+ © ESO 2021
4
+
5
+ A large sub-Neptune transiting the thick-disk M4 V TOI-2406
6
+ , B. V. Rackham2,3,(cid:63), N. Schanche1, R. Petrucci4,5, Y. Gómez Maqueo Chew6, B.-O. Demory1,
7
+ R. D. Wells1
8
+
9
+ A. J. Burgasser7, R. Burn8, F. J. Pozuelos9,10, M. N. Günther3,11,(cid:63)(cid:63), L. Sabin12, U. Schroffenegger1,
10
+ M. A. Gómez-Muñoz12, K. G. Stassun13, V. Van Grootel10, S. B. Howell14, D. Sebastian15, A. H. M. J. Triaud15,
11
+ D. Apai16,17,(cid:63)(cid:63)(cid:63), I. Plauchu-Frayn12, C. A. Guerrero12, P. F. Guillén12, A. Landa12, G. Melgoza12, F. Montalvo12,
12
+ H. Serrano12, H. Riesgo12, K. Barkaoui9,18, A. Bixel16, A. Burdanov2, W. P. Chen19, P. Chinchilla9,20, K. A. Collins21,
13
+ T. Daylan3,11,(cid:63)(cid:63)(cid:63)(cid:63), J. de Wit2, L. Delrez9,10, M. Dévora-Pajares22, J. Dietrich16, G. Dransfield15, E. Ducrot9,
14
+ M. Fausnaugh3,11, E. Furlan23, P. Gabor24, T. Gan25, L. Garcia9, M. Ghachoui18, S. Giacalone26, A. B. Gibbs27,
15
+ M. Gillon9, C. Gnilka14, R. Gore26, N. Guerrero3,11, T. Henning8, K. Hesse3,11, E. Jehin10, J. M. Jenkins14,
16
+ D. W. Latham21, K. Lester14, J. McCormac28, C. A. Murray29, P. Niraula2, P. P. Pedersen29, D. Queloz29, G. Ricker3,11,
17
+ D. R. Rodriguez30, A. Schroeder26, R. P. Schwarz31, N. Scott14, S. Seager2,3,11,32, C. A. Theissen7,(cid:63)(cid:63)(cid:63)(cid:63)(cid:63), S. Thompson29,
18
+ M. Timmermans9, J. D. Twicken33,14, and J. N. Winn34
19
+
20
+ (Affiliations can be found after the references)
21
+
22
+ Received 10 May 2021 / Accepted 23 July 2021
23
+
24
+ ABSTRACT
25
+
26
+ Context. Large sub-Neptunes are uncommon around the coolest stars in the Galaxy and are rarer still around those that are metal-
27
+ poor. However, owing to the large planet-to-star radius ratio, these planets are highly suitable for atmospheric study via transmission
28
+ spectroscopy in the infrared, such as with JWST.
29
+ Aims. Here we report the discovery and validation of a sub-Neptune orbiting the thick-disk, mid-M dwarf star TOI-2406. The star’s
30
+ low metallicity and the relatively large size and short period of the planet make TOI-2406 b an unusual outcome of planet formation,
31
+ and its characterisation provides an important observational constraint for formation models.
32
+ Methods. We first infer properties of the host star by analysing the star’s near-infrared spectrum, spectral energy distribution, and
33
+ Gaia parallax. We use multi-band photometry to confirm that the transit event is on-target and achromatic, and we statistically validate
34
+ the TESS signal as a transiting exoplanet. We then determine physical properties of the planet through global transit modelling of the
35
+ TESS and ground-based time-series data.
36
+ Results. We determine the host to be a metal-poor M4 V star, located at a distance of 56 pc, with properties Teff = 3100 ± 75 K,
37
+ M∗ = 0.162 ± 0.008 M(cid:12), R∗ = 0.202 ± 0.011 R(cid:12), and [Fe/H] = −0.38 ± 0.07, and a member of the thick disk. The planet is a relatively
38
+ large sub-Neptune for the M-dwarf planet population, with Rp = 2.94 ± 0.17 R⊕ and P = 3.077 d, producing transits of 2% depth. We
39
+ note the orbit has a non-zero eccentricity to 3σ, prompting questions about the dynamical history of the system.
40
+ Conclusions. This system is an interesting outcome of planet formation and presents a benchmark for large-planet formation around
41
+ metal-poor, low-mass stars. The system warrants further study, in particular radial velocity follow-up to determine the planet mass and
42
+ constrain possible bound companions. Furthermore, TOI-2406 b is a good target for future atmospheric study through transmission
43
+ spectroscopy. Although the planet’s mass remains to be constrained, we estimate the S/N using a mass-radius relationship, ranking
44
+ the system fifth in the population of large sub-Neptunes, with TOI-2406 b having a much lower equilibrium temperature than other
45
+ spectroscopically accessible members of this population.
46
+
47
+ Key words. planets and satellites: detection – stars: individual: TOI-2406 – techniques: photometric
48
+
49
+ 1. Introduction
50
+
51
+ Short-period, rocky planets have been found to be common
52
+ around M dwarfs; however, larger planets are rarer (Dressing
53
+ & Charbonneau 2015; Mulders et al. 2015; Hardegree-Ullman
54
+ et al. 2019). This extends down to the coolest stars, where only
55
+ seven transiting large sub-Neptune (2.75 < Rp < 4.0 R⊕) plan-
56
+ ets are known to orbit stars with effective temperatures below
57
+
58
+ (cid:63) 51 Pegasi b Fellow.
59
+ (cid:63)(cid:63) Juan Carlos Torres Fellow.
60
+ (cid:63)(cid:63)(cid:63) NASA Nexus For Exoplanetary System Science: Alien Earths
61
+
62
+ Team.
63
+
64
+ (cid:63)(cid:63)(cid:63)(cid:63) Kavli Fellow.
65
+ (cid:63)(cid:63)(cid:63)(cid:63)(cid:63) NASA Sagan Fellow.
66
+
67
+ 3300 K1. The search for planets around these cool stars is ongo-
68
+ ing with surveys such as SPECULOOS (Delrez et al. 2018a;
69
+ Sebastian et al. 2020), MEarth (Nutzman & Charbonneau 2008),
70
+ CARMENES (Reiners et al. 2018; Morales et al. 2019) and
71
+ EDEN (Gibbs et al. 2020).
72
+
73
+ From core accretion models (Pollack et al. 1996), the occur-
74
+ rence rate of large sub-Neptunes around late-M dwarfs is
75
+ expected to be lower for metal-deficient environments (Burn
76
+ et al. 2021). Therefore, if stellar metallicity is a good proxy for
77
+ the amount of solid material in the protoplanetary disk, we would
78
+ not expect to readily find large sub-Neptunes around metal-poor
79
+ stars. In particular, planetesimal-based formation models do not
80
+ readily produce these systems, and therefore their existence can
81
+
82
+ 1 NASA Exoplanet Archive (04 March 2021).
83
+
84
+ Article published by EDP Sciences
85
+
86
+ A97, page 1 of 17
87
+
88
+ Astronomy&Astrophysics A&A 653, A97 (2021)
89
+
90
+ Fig. 1. Target pixel file (TPF) images for TOI-2406, from TESS Sectors 3 (left) and 30 (right). The apertures used for light curve generation are
91
+ over-plotted; sources identified in Gaia DR2 are also included, with symbols correlated to their brightness compared to the target. These images
92
+ were produced using tpfplotter (Aller et al. 2020).
93
+
94
+ be better explained by a pebble accretion scenario, in which the
95
+ available solids can be transported to the planet from the outer
96
+ disk.
97
+
98
+ The large planet-to-star radius ratios for planets around cool
99
+ stars means they produce deep transits. This makes them good
100
+ targets for atmospheric study via transmission spectroscopy
101
+ (Seager & Sasselov 2000; Brown 2001), particularly in the
102
+ infrared, where M dwarfs are relatively bright. These wave-
103
+ lengths can be probed with HST and soon JWST. Notably,
104
+ GJ 3470 b (Bonfils et al. 2012; Teff = 3600 K, Rp = 4.6 R⊕)
105
+ and GJ 1214 b (Charbonneau et al. 2009; Teff = 3026 K, Rp =
106
+ 2.8 R⊕) are two of the most well-studied planets by transmission
107
+ spectroscopy1.
108
+
109
+ Here we report the discovery and validation of a sub-Neptune
110
+ planet orbiting a metal-poor M4 V host star (Teff = 3100 K, Rp =
111
+ 2.9 R⊕). In Sect. 2 we detail the TESS and follow-up observa-
112
+ tions. In Sect. 3 we present the stellar characterisation analysis.
113
+ Then in Sect. 4, owing to improved precision over TESS from the
114
+ ground, we validate the transit signal as a bona fide planet. We
115
+ present our search for further planets in the TESS data in Sect. 5,
116
+ before giving the results from our transit modelling of the pho-
117
+ tometric light curves in Sect. 6. Finally, in Sect. 7 we discuss
118
+ the system’s formation, architecture, and potential for follow-up
119
+ study, and we conclude in Sect. 8.
120
+
121
+ 2. Observations
122
+
123
+ 2.1. TESS photometry
124
+
125
+ TOI-2406 (TIC 212957629) was first observed by TESS (Ricker
126
+ et al. 2015) in the full frame images (FFIs) at 30-min cadence
127
+ in Sector 3, between 20 September and 18 October 2018. It was
128
+ observed by TESS again two years later, at 2-min cadence in
129
+ Sector 30, between 22 September and 21 October 2020. The
130
+ two sectors are shown in Fig. 1, along with the nearby Gaia
131
+ sources (Gaia Collaboration 2018). The star is present in the
132
+ TESS Input Catalogue v8.1 (TIC; Stassun et al. 2019), identi-
133
+ fied as an M4.5 dwarf located at 56 pc. The Science Processing
134
+ Operations Center (SPOC; Jenkins et al. 2016) at NASA Ames
135
+ Research Center extracted the photometry for this target from
136
+ the 2-min data from Sector 30, and conducted a transiting planet
137
+ search (Jenkins 2002; Jenkins et al. 2010) on 30 October 2020.
138
+
139
+ A97, page 2 of 17
140
+
141
+ This yielded a strong 2% deep transit-like signature at a period
142
+ of 3.077 d. The TESS Object of Interest (TOI) vetting team at
143
+ MIT reviewed the SPOC Data Validation reports (Twicken et al.
144
+ 2018; Li et al. 2019) for this signal and released TOI-2406 on 24
145
+ November 2020 (Guerrero et al. 2021).
146
+
147
+ We retrieved the TESS Sector 3 FFI light curve via point
148
+ spread function (PSF) photometry on a 11 × 11 pixel region
149
+ around the target, which we downloaded via TESSCut (Brasseur
150
+ et al. 2019). In order to perform photometry, we retrieved
151
+ the source position from TICv8 and corrected it for proper
152
+ motion. We then interpolated the empirical, super-sampled pixel
153
+ response function2 (PRF) using a third-order polynomial. A flat
154
+ template for the background sky emission was used. Finally, we
155
+ determined the fluxes (i.e. template coefficients) of the target and
156
+ the sky background for all time bins via linear regression. The
157
+ data release notes for Sector 33 indicate that TESS conducted a
158
+ series of engineering tests during its usual data gap (the perigee
159
+ orbit) to improve the pointing accuracy. We inspected the data to
160
+ confirm that its quality is reliable and included it in our analysis.
161
+ We downloaded the TESS Sector 30 SPOC Presearch Data
162
+ Conditioning Simple Aperture Photometry (PDC-SAP; Stumpe
163
+ et al. 2012, 2014; Smith et al. 2012) light curve from MAST.
164
+ Long-term trends are already removed in the PDC light curve.
165
+ We assessed further detrending via a median filter but found
166
+ no improvement to the light curve quality, likely because it is
167
+ dominated by white noise. All data with a non-zero quality flags
168
+ were removed, along with the final ∼1 d of the sector. We also
169
+ tested removing the first transit of the sector, but found no change
170
+ in our Markov chain Monte Carlo (MCMC) results, and there-
171
+ fore included these data in our analysis. Both TESS light curves
172
+ are displayed in Fig. 2, showing the transit times and regions
173
+ excluded from the transit analysis.
174
+
175
+ 2.2. Follow-up photometry
176
+
177
+ We obtained ground-based follow-up of multiple transit events,
178
+ coordinated by the TESS Follow-up Observing Program (TFOP)
179
+
180
+ 2 https://archive.stsci.edu/tess/all_products.html
181
+ 3 https://archive.stsci.edu/missions/tess/doc/tess_
182
+ drn/tess_sector_03_drn04_v02.pdf
183
+
184
+ 698700702704706708710Pixel Column Number110112114116118120122Pixel Row NumberENm= -2m= 0m= 2m= 4m= 6123456789101112TIC 212957629 - Sector 32.02.53.03.54.04.5Flux ×102 (e)122612281230123212341236Pixel Column Number110112114116118120Pixel Row NumberENm= -2m= 0m= 2m= 4m= 6123456789101112TIC 212957629 - Sector 300.20.40.60.81.01.21.41.6Flux ×102 (e) R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
185
+
186
+ Fig. 2. Sector 3 light curve created from FFIs (top) and the Sector 30 2-min light curve (bottom). The 2-min data points (grey) have been binned
187
+ by 15 to produce the black points, with error bars corresponding to the standard deviation in the bins, which match the FFI light curve cadence
188
+ of 30 min. Regions marked in red were excluded from the analysis: in Sector 3, there were poor-quality engineering data between orbits; and in
189
+ Sector 30, a significant trend at the end was possibly caused by scattered light. Regions marked in grey correspond to data taken outside of science
190
+ data acquisition for testing purposes but used in the analysis. One transit occurred in this period, towards the end of Sector 3. The transits of
191
+ TOI-2406 b are labelled with green markers below the light curves.
192
+
193
+ Table 1. Summary of photometric observations.
194
+
195
+ Night of (UT)
196
+
197
+ 20 Sep.–18 Oct. 2018
198
+ 22 Sep.–21 Oct. 2020
199
+ 29 Nov. 2020
200
+ 02 Dec. 2020
201
+ 05 Jan. 2021
202
+ 05 Jan. 2021
203
+ 09 Jan. 2021
204
+
205
+ Ntr
206
+
207
+ 6
208
+ 8
209
+ 1
210
+ 1
211
+ <1
212
+ 1
213
+ 1
214
+
215
+ Telescope
216
+
217
+ Filter
218
+
219
+ Exposure (s) σ30min (ppt)
220
+
221
+ TESS Sector 3
222
+ TESS Sector 30
223
+ TRAPPIST-South
224
+ SAINT-EX
225
+ TRAPPIST-South
226
+ EDEN/VATT
227
+ LCO-McDonald
228
+
229
+ TESS
230
+ TESS
231
+ Exo
232
+ z(cid:48)
233
+ Exo
234
+ V
235
+ i(cid:48)
236
+
237
+ 1800
238
+ 120
239
+ 120
240
+ 120
241
+ 120
242
+ 120
243
+ 150
244
+
245
+ 6.5
246
+ 5.7
247
+ 1.4
248
+ 0.5
249
+ –
250
+ 2.6
251
+ 0.9
252
+
253
+ Notes. Shown are: the date(s) of the observations; the number of transits covered; the telescope, filter, and exposure time used; and the RMS of the
254
+ 30-min binned light curve.
255
+
256
+ sub-group 1 for Seeing-limited Photometry. The observations are
257
+ summarised in Table 1 and detailed in the following sections.
258
+
259
+ 2.2.1. TRAPPIST-South
260
+
261
+ TOI-2406 was first followed-up on 29 November 2020 with the
262
+ 0.6 m TRAPPIST-South telescope, located at La Silla Observa-
263
+ tory, Chile (Jehin et al. 2011; Gillon et al. 2011). A full transit
264
+ was observed in the Exo bandpass (500–1100 nm), consisting of
265
+ 74 images each with an exposure time of 120 s, covering 168 min.
266
+ The data were reduced, and a light curve was produced using the
267
+ AstroImageJ (AIJ) software (Collins et al. 2017). A partial tran-
268
+ sit was also observed on 05 January 2021; however, this was not
269
+ used in our analysis due to the small amount of transit coverage.
270
+
271
+ 2.2.2. SAINT-EX
272
+
273
+ We obtained a full transit with SAINT-EX on 02 December 2020
274
+ in z(cid:48). SAINT-EX (Search And characterIsatioN of Transiting
275
+ EXoplanets) is a 1-m telescope located at the Observatorio
276
+
277
+ Astronómico Nacional, San Pedro Mártir, Mexico (Sabin et al.
278
+ 2018). The observations consisted of 79 images with an expo-
279
+ sure time of 120 s, covering 178 min total. The data were reduced
280
+ using a custom pipeline, PRINCE, which is detailed in Demory
281
+ et al. (2020). The final light curve was produced by a weighted
282
+ principal component analysis (PCA) approach (Bailey 2012)4 to
283
+ correct for common trends between the target and comparison
284
+ stars. We also produced a light curve using AIJ, but found it
285
+ to have a higher out-of-transit scatter and a larger trend. We
286
+ therefore used the PRINCE light curve in our analysis. The
287
+ SAINT-EX light curve had a 30-min. rms of ∼500 ppm (see
288
+ Table 1), with a good amount of baseline on either side of transit,
289
+ making this our most constraining dataset.
290
+
291
+ 2.2.3. VATT
292
+
293
+ We also obtained a V-band transit with the VATT4K CCD
294
+ Imager on the 1.8 m Vatican Advanced Technology Telescope
295
+
296
+ 4 https://github.com/jakevdp/wpca
297
+
298
+ A97, page 3 of 17
299
+
300
+ 838583908395840084050.920.961.001.041.08Normalised fluxSector 3 30-min FFI91209125913091359140Time (BJD - 2450000)0.920.961.001.041.08Normalised fluxSector 30 2-min A&A 653, A97 (2021)
301
+
302
+ (VATT) on 06 January 2021 (UT). The data were collected as
303
+ part of Project EDEN (Exo-Earth Discovery and Exploration
304
+ Network; e.g. Gibbs et al. 2020). We collected 88 images each
305
+ with an exposure time of 120 s over 201 min, covering almost
306
+ the full transit. We initially followed the data reduction and
307
+ calibration procedure for Project EDEN observations, outlined
308
+ by Gibbs et al. (2020). We then also produced a light curve
309
+ using AIJ for aperture photometry and weighted PCA for dif-
310
+ ferential photometry, which we found to be improved over the
311
+ initial pipeline reduction, and therefore we used this in our
312
+ analysis.
313
+
314
+ 2.2.4. Las Cumbres Observatory (LCO)
315
+
316
+ A final transit was observed on 09 January 2021 with the Las
317
+ Cumbres Observatory Global Telescope (LCOGT; Brown et al.
318
+ 2013). The observations were taken using the 1-m telescope at
319
+ McDonald Observatory, using the i(cid:48) filter. We took 55 exposures
320
+ of 150 s over 173 min, covering the full transit, although lacking
321
+ some post-transit baseline. The data were reduced, and a light
322
+ curve was produced using AIJ.
323
+
324
+ Fig. 3. Zorro speckle imaging 5σ contrast curves, along with the
325
+ reconstructed 832 nm image.
326
+
327
+ 2.3. Spectroscopy
328
+
329
+ 3. Stellar properties
330
+
331
+ Infrared spectroscopy of TOI-2406 was obtained with the SpeX
332
+ spectrograph (Rayner et al. 2003) on the 3.2-m NASA Infrared
333
+ Telescope Facility on Maunakea, Hawaii, on 03 December
334
+ 2020 (UT). Conditions were clear with 0.(cid:48)(cid:48)4 seeing. The short-
335
+ wavelength cross-dispersed (SXD) mode was used with the
336
+ 0.(cid:48)(cid:48)3-wide slit to obtain a 0.7–2.5 µm spectrum in seven orders at
337
+ a spectral resolving power λ/∆λ ≈ 2000. A total of four ABBA
338
+ nod sequences (16 exposures) were obtained with an integra-
339
+ tion time of 170 s per exposure with the slit aligned with the
340
+ parallactic angle. The A0 V star HD 13936 (V = 6.549) was
341
+ observed afterwards at an equivalent airmass for flux and telluric
342
+ calibration, followed by arc lamp and flat field lamp exposures.
343
+ Data were reduced using SpeXtool v4.1 (Cushing et al. 2004)
344
+ using standard settings. The resulting spectrum of TOI-2406 had
345
+ a median signal-to noise ratio (S/N) of 60, with JHK peaks of
346
+ around 100–150.
347
+
348
+ 2.4. High-resolution imaging
349
+
350
+ TOI-2406 was observed with the Zorro instrument mounted
351
+ on the 8-metre Gemini-South telescope on 29 December 2020.
352
+ Zorro simultaneously observes in blue (562/54 nm) and red
353
+ (832/40 nm) bandpasses, with inner working angles of 0.(cid:48)(cid:48)026 in
354
+ the blue and 0.(cid:48)(cid:48)017 in the red. Three thousand 0.06-s images
355
+ were obtained and combined in the data reduction process and
356
+ the Fourier analysis as described in Howell et al. (2011). The
357
+ Zorro observations do not reveal a previously unknown com-
358
+ panion to TOI-2406 within the 5σ contrast limits obtained (see
359
+ Fig. 3). The contrast curve begins to flatten at ∼0.(cid:48)(cid:48)1 and appears
360
+ flat from ∼0.(cid:48)(cid:48)2. Given the distance to the target of 55.6 pc, these
361
+ correspond to projected distances of 5.6 and 11.1 AU away from
362
+ the star, respectively. The observations had a contrast of 4–
363
+ 5 magnitudes beyond 0.(cid:48)(cid:48)2, allowing us to rule out companion
364
+ stars earlier than about spectral type M8 beyond this separa-
365
+ tion. We also find no evidence for companion stars within 0.(cid:48)(cid:48)2
366
+ in the near-infrared spectrum. Therefore, any possible compan-
367
+ ions would be unable to explain the 2% transit depth seen in the
368
+ TESS and follow-up data.
369
+
370
+ A97, page 4 of 17
371
+
372
+ TOI-2406 is a quiet, mid-M dwarf located towards the eclip-
373
+ tic plane (l = −6.6◦) at 56 pc. As we show below from an
374
+ analysis of the space velocity, it is very likely a member of
375
+ the Galactic thick-disk population, and therefore would have an
376
+ age of 11 ± 1.5 Gyr (Miglio et al. 2021). The star is faint in
377
+ the optical (V (cid:38) 17) but reasonably bright in the near-infrared
378
+ (J = 12.6). The catalogued astrometric and photometric param-
379
+ eters of TOI-2406 are given in Table 2. In this section we detail
380
+ the methodology used to determine the properties of the star,
381
+ given in Table 3.
382
+
383
+ 3.1. Spectroscopy
384
+
385
+ The near-infrared spectrum of TOI-2406 is shown in Fig. 4
386
+ in comparison to equivalent data for the M4 spectral standard
387
+ Gliese 213 from the IRTF Spectral Library (Rayner et al. 2009).
388
+ Both the overall infrared spectral energy distributions (SEDs)
389
+ and the detailed line features in these spectra are well matched.
390
+ We evaluated the metallicity-sensitive lines Na I (2.2079 µm)
391
+ and Ca I (2.2640 µm) using the empirical calibrations of Mann
392
+ et al. (2013) to infer an average metallicity of [Fe/H] = −0.38 ±
393
+ 0.07 dex (i.e. significantly metal-poor compared to the Sun). We
394
+ also used the line centres for Na I, Mg I, Al I, K I, and Ca I to
395
+ infer a heliocentric velocity of +15±6 km s−1. Combining this
396
+ with the tangential velocity from Gaia Early Data Release 3
397
+ (EDR3; Gaia Collaboration 2021), we infer the local standard
398
+ of rest (LSR) UVW velocities listed in Table 2.
399
+
400
+ The significant negative V velocity relative to the LSR makes
401
+ this source a likely thick-disk star, based on the kinematic sam-
402
+ ple of Bensby et al. (2003). Following the methodology of
403
+ Bensby et al. (2014), we compute membership probabilities of
404
+ 0.3, 99.5, and 0.2 per cent for the thin-disk, thick-disk, and
405
+ halo populations, respectively. Therefore, we reason that TOI-
406
+ 2406 is a member of the thick disk, which is consistent with the
407
+ inferred sub-solar metallicity. We note that Gliese 213 is also a
408
+ high-velocity star with thick-disk kinematics and a slightly sub-
409
+ solar metallicity based on its near-infrared spectrum ([Fe/H] =
410
+ −0.27±0.05 dex).
411
+
412
+ R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
413
+
414
+ Table 2. Astrometric and photometric properties of the host star.
415
+
416
+ Property
417
+
418
+ Identifiers:
419
+ TOI
420
+ TIC
421
+ LP
422
+ 2MASS
423
+ Gaia ID
424
+
425
+ Astrometry:
426
+ RA (J2000)
427
+ Dec (J2000)
428
+ µRA (mas yr−1)
429
+ µDec (mas yr−1)
430
+ Parallax (mas)
431
+ Distance (pc)
432
+ Vtan (km s−1)
433
+ RV (km s−1)
434
+ U (km s−1)
435
+ V (km s−1)
436
+ W (km s−1)
437
+
438
+ Photometry:
439
+ TESS (mag)
440
+ BP (mag)
441
+ G (mag)
442
+ RP (mag)
443
+ g (mag)
444
+ r (mag)
445
+ i (mag)
446
+ z (mag)
447
+ y (mag)
448
+ J (mag)
449
+ H (mag)
450
+ K (mag)
451
+ W1 (mag)
452
+ W2 (mag)
453
+ W3 (mag)
454
+ W4 (mag)
455
+
456
+ Value
457
+
458
+ Source
459
+
460
+ 2406
461
+ 212957629
462
+ 645-50
463
+ J00351318-0322140
464
+ 2528453161326406016
465
+
466
+ 00:35:13.22
467
+ −03:22:14.29
468
+ 226.01 ± 0.05
469
+ −336.24 ± 0.04
470
+ 17.98 ± 0.041
471
+ 55.60 ± 0.13
472
+ 106.8 ± 0.2
473
+ +15 ± 6
474
+ +1.5 ± 0.9
475
+ −93.2 ± 2.1
476
+ −46.4 ± 5.2
477
+
478
+ 14.31
479
+ 17.434 ± 0.007
480
+ 15.663 ± 0.003
481
+ 14.393 ± 0.004
482
+ 17.7198 ± 0.0081
483
+ 16.5656 ± 0.0037
484
+ 14.9479 ± 0.0084
485
+ 14.2083 ± 0.0037
486
+ 13.8748 ± 0.0049
487
+ 12.633 ± 0.024
488
+ 12.129 ± 0.024
489
+ 11.894 ± 0.025
490
+ 11.706 ± 0.023
491
+ 11.458 ± 0.022
492
+ 11.049 ± 0.143
493
+ >8.593
494
+
495
+ [1]
496
+ [1]
497
+ [1]
498
+ [1]
499
+ [1]
500
+ [1]
501
+ [1]
502
+ [2]
503
+ [2]
504
+ [2]
505
+ [2]
506
+
507
+ [3]
508
+ [1]
509
+ [1]
510
+ [1]
511
+ [4]
512
+ [4]
513
+ [4]
514
+ [4]
515
+ [4]
516
+ [5]
517
+ [5]
518
+ [5]
519
+ [6]
520
+ [6]
521
+ [6]
522
+ [6]
523
+
524
+ References. (1) Gaia EDR3 (Gaia Collaboration 2021); (2) this work;
525
+ (3) TIC (Stassun et al. 2019); (4) Pan-STARRS (Chambers et al. 2016);
526
+ (5) 2MASS (Skrutskie et al. 2006); (6) WISE (Cutri et al. 2021).
527
+
528
+ Table 3. Properties of the host star.
529
+
530
+ Property
531
+
532
+ Value
533
+
534
+ Source
535
+
536
+ Sp. type
537
+ Teff (K)
538
+ [Fe/H]
539
+ M∗ (M(cid:12))
540
+ R∗ (R(cid:12))
541
+ Fbol (10−11 erg s−1 cm−2)
542
+ L∗ (10−3 L(cid:12))
543
+ log g
544
+ ρ∗ (g cm−3)
545
+
546
+ M4
547
+ 3100 ± 75
548
+ −0.38 ± 0.07
549
+ 0.162 ± 0.008
550
+ 0.202 ± 0.011
551
+ 3.53 ± 0.17
552
+ 3.40 ± 0.16
553
+ 5.037+0.053
554
+ −0.051
555
+ 27.7+5.3
556
+ −4.2
557
+
558
+ Spectrum
559
+ SED
560
+ Spectrum
561
+ See text
562
+ SED
563
+ SED
564
+ SED
565
+ M∗, R∗
566
+ M∗, R∗
567
+
568
+ Notes. Metallicity is based on the empirical calibrations of Mann et al.
569
+ (2014) and Newton et al. (2014).
570
+
571
+ Fig. 4. SpeX/SXD near-infrared spectrum of TOI-2406 (black line)
572
+ compared to equivalent data for the M4.0 spectral standard Gliese
573
+ 213 (data from Rayner et al. 2009; magenta line). Top: full spectra,
574
+ with the uncertainty spectrum of TOI-2406 is shown in black along
575
+ the bottom, and regions of strong telluric absorption are indicated by
576
+ the grey panels. Bottom: close-up of the K-band region for these two
577
+ spectra, highlighting the metallicity-sensitive absorption features Na I
578
+ (λ2.2079 µm) and Ca I (λ2.2640 µm), as well as CO band heads.
579
+ The difference between the spectra (blue line) is consistent with the
580
+ measurement uncertainties (grey band).
581
+
582
+ 3.2. SED fitting, empirical relations, and evolutionary
583
+
584
+ modelling
585
+
586
+ We performed an analysis of the broadband SED of the star
587
+ together with the Gaia EDR3 parallax (with no systematic offset
588
+ applied; see e.g. Stassun & Torres 2021), in order to determine an
589
+ empirical measurement of the stellar radius, following the proce-
590
+ dures described in Stassun & Torres (2016); Stassun et al. (2017,
591
+ 2018). We pulled the grizy magnitudes from Pan-STARRS, the
592
+ JHKS magnitudes from 2MASS, and the W1–W3 magnitudes
593
+ from WISE. We opted not to use the G, GBP, GRP magnitudes
594
+ from Gaia, as these very broad filters are less ideal than the
595
+ narrower bandpasses that are available. Together, the available
596
+ photometry spans the stellar SED over the wavelength range
597
+ 0.4–10 µm (see Fig. 5).
598
+
599
+ We performed a fit using the BT-Dusty stellar atmo-
600
+ sphere models (Allard et al. 2012), fitting for the effective
601
+ temperature (Teff), metallicity ([Fe/H]), and surface gravity
602
+ (log g). The extinction (AV) was fixed at zero due to the
603
+ star’s proximity. The resulting fit, with Teff = 3100 ± 75 K,
604
+ log g = 5.0 ± 0.5, and [Fe/H] = −0.5 ± 0.5, is reasonably good
605
+ (Fig. 5), with a reduced χ2 of 1.4. Integrating the (un-reddened)
606
+ model SED gives the bolometric flux at Earth, Fbol = 3.53 ±
607
+ 0.17 × 10−11 erg s−1 cm−2. Taking the Fbol and Teff, together
608
+
609
+ A97, page 5 of 17
610
+
611
+ 0.81.01.21.41.61.82.02.22.4Wavelength (m)0.00.20.40.60.81.0Normalized Flux (F)TOI-2406Gliese 213 (M4)2.182.202.222.242.262.282.302.322.34Wavelength (m)0.50.60.70.80.91.01.11.2Normalized Flux (F)Na ICa ICOCOTOI-2406Gliese 213 (M4)Difference A&A 653, A97 (2021)
612
+
613
+ 0.6σ, and there was no apparent secondary eclipse. TESS uses
614
+ a large spatial pixel scale of 20(cid:48)(cid:48), making the aperture of a
615
+ single star ∼1(cid:48), which could be contaminated by light from
616
+ nearby stars. The star has no resolved neighbours within
617
+ 1 arcmin, and those nearest are faint and do not provide enough
618
+ light in the aperture to produce the observed 2% deep signal. All
619
+ other tests, including centroid offset and ghost diagnostic, were
620
+ also passed.
621
+
622
+ 4.2. Follow-up photometry
623
+
624
+ The ground-based follow-up photometry allowed a two-fold test
625
+ of false-positive scenarios: first, by confirming the TESS signal
626
+ was on the expected star, and second, by assessing the transit
627
+ depth at multiple wavelengths. The follow-up photometry also
628
+ has much-improved precision over the TESS data, in particular
629
+ the SAINT-EX light curve, allowing a more stringent test of the
630
+ transit shape (see Sect. 4.4).
631
+
632
+ In contrast to the TESS large pixel scale, the follow-up
633
+ photometry we obtained were extracted with apertures of a
634
+ few arcseconds. No dimming events were seen on any nearby
635
+ stars, while we detected clear 2% transits on the target in each
636
+ follow-up light curve.
637
+
638
+ Our photometric observations comprised five different band-
639
+ passes – TESS, V, i(cid:48), z(cid:48), and Exo – covering a wavelength range
640
+ of approximately 500–1100 nm. This allowed us to test the chro-
641
+ maticity of the transit by measuring the depth individually in
642
+ each band. We find the depths to agree to better than 1.5σ, and
643
+ no trend is seen with wavelength.
644
+
645
+ 4.3. Archival imaging
646
+
647
+ We inspected archival
648
+ images spanning 69 years to con-
649
+ strain the presence of an astrophysical false positive signal
650
+ at the current position of TOI-2406. Given its high proper
651
+ motion (405 mas yr−1, Gaia Collaboration 2018), images taken
652
+ as recently as 2008 are useful for this purpose. We also note that
653
+ the nearest neighbour in the Gaia catalogue is at a distance of
654
+ 65(cid:48)(cid:48)and the re-normalised unit weight error (RUWE) is 1.06; that
655
+ is to say, the target has no close neighbours and its astrometry is
656
+ explained well by a single-star model (Belokurov et al. 2020).
657
+ In total, we considered POSS I/DSS (Minkowski & Abell 1963;
658
+ Lasker et al. 1990), POSS II/DSS2 (Reid et al. 1991), 2MASS
659
+ (Skrutskie et al. 2006), and SDSS-III (Alam et al. 2015) imagery,
660
+ none of which show any point sources at the current position of
661
+ the target. Figure 6 shows a subset of these images.
662
+
663
+ The SDSS i and z images provide the tightest constraints on
664
+ background sources. The target is offset in these images by 4.(cid:48)(cid:48)9
665
+ from its current position. The reported magnitudes of the target
666
+ are 14.92 and 14.08 in i and z, respectively, and no other objects
667
+ at the current position of the target are detected. Given the mag-
668
+ nitude limits of SDSS DR126, both of these images allow us to
669
+ rule out sources at the current position of the target that are 6
670
+ .4 mag (360 times) fainter than TOI-2406. Thus, an undetected
671
+ background object could contribute no more than 0.3% of the
672
+ total flux we observe on-target. Even the worst-case scenario – a
673
+ 100% drop in brightness on an object at the magnitude limit of
674
+ the SDSS images – would not account for the 2% transit feature
675
+ we observe. We conclude that the transit signal originates from
676
+ the TOI-2406 system and not from another astrophysical source
677
+ along the same line of sight.
678
+
679
+ 6 https://www.sdss.org/dr12/scope/
680
+
681
+ Fig. 5. SED fit of TOI-2406. The black curve is the best fitting BT-Dusty
682
+ model, red symbols are the observed fluxes (horizontal bars represent
683
+ the effective bandpass widths), and blue symbols are the model fluxes.
684
+
685
+ with the Gaia parallax, gives the stellar radius, R(cid:63) = 0.202 ±
686
+ 0.011 R(cid:12).
687
+
688
+ To assess the reliability of our stellar radius measurement,
689
+ we compared our SED model result to empirical relations for
690
+ cool stars. Both the Mann et al. (2015) MK−R∗ relation and the
691
+ Boyajian et al. (2012) M∗−R∗ relation (see discussion of M∗
692
+ in the following paragraphs) return a stellar radius of 0.197 ±
693
+ 0.010 R(cid:12). This is in good agreement with our SED radius (0.5σ),
694
+ which we use as our final stellar radius estimate.
695
+
696
+ For the mass, we applied stellar evolutionary modelling,
697
+ using the models for very low-mass stars presented in Fernandes
698
+ et al. (2019). We used as constraints the luminosity derived from
699
+ Fbol and the Gaia EDR3 parallax (L(cid:63) = 3.44 ± 0.16 × 10−3L(cid:12)),
700
+ the metallicity inferred in Sect. 3.1, and assuming an age of
701
+ (cid:38)2 Gyr, in the absence of signs of a young star (e.g. no fast rota-
702
+ tion and no flares seen in the light curves; Sect. 2)5. We obtained
703
+ a stellar mass of 0.163 ± 0.007 M(cid:12). This uncertainty reflects
704
+ the error propagation on the stellar luminosity and metallicity,
705
+ but also the uncertainty associated with the input physics of the
706
+ stellar models (Van Grootel et al. 2018).
707
+
708
+ We also estimated the mass through the empirical relations,
709
+ the mass-MK relation of Mann et al. (2019) and the mass-
710
+ luminosity relation of Benedict et al. (2016), finding values of
711
+ 0.165 ± 0.008 and 0.158 ± 0.008 M∗, respectively. Combining
712
+ these empirical estimates and our evolutionary modelling value
713
+ by a simple average, we obtain our final stellar mass estimate
714
+ of 0.162 ± 0.008 M(cid:12). Considering the stellar radius estimate
715
+ from SED fitting, this gives a mean stellar density of ρ(cid:63) =
716
+ 27.7+5.3
717
+
718
+ −4.2 g cm−3.
719
+
720
+ 4. Planet validation
721
+
722
+ 4.1. TESS data validation report
723
+
724
+ We initially vetted the target before obtaining further observa-
725
+ tions by examining the TESS Data Validation Report for Sector
726
+ 30 (Twicken et al. 2018; Li et al. 2019). TOI-2406.01 was sum-
727
+ marised as a strong periodic transit signal with a phase-folded
728
+ S/N of 13.2. The odd and even transit depths agreed to within
729
+
730
+ 5 The luminosity of very low-mass stars evolves very slowly with time
731
+ once the star has turned on core H-burning and has reached the main
732
+ sequence.
733
+
734
+ A97, page 6 of 17
735
+
736
+ 110(cid:104) (µm)-12.5-12.0-11.5-11.0-10.5log (cid:104)F(cid:104) (erg s-1 cm-2) R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
737
+
738
+ Fig. 6. Imaging of the current position of TOI-2406, spanning seven decades. Archival images from POSS I, POSS II, and SDSS are shown, along
739
+ with the median image from the SAINT-EX observations. For each image, the bandpass and epoch are noted, and the position of TOI-2406 during
740
+ the SAINT-EX observations is highlighted. The archival images were retrieved using the astroquery (Ginsburg et al. 2019) interface to SkyView7.
741
+
742
+ 4.4. False-positive likelihood
743
+
744
+ 5. Planet searches and detection limits from the
745
+
746
+ To fully vet the planet candidate, we utilised two open-source
747
+ software packages, triceratops (Giacalone & Dressing 2020;
748
+ Giacalone et al. 2021) and vespa (Morton 2012, 2015b). Both of
749
+ these simulate various false-positive scenarios, allowing a cal-
750
+ culation of the likelihood that the transit signal is caused by a
751
+ planet. To statistically validate the planet, we require a false pos-
752
+ itive probability (FPP) less than 0.01 (1%), as is typically used
753
+ in the field.
754
+
755
+ The triceratops tool was developed to aid in the vetting
756
+ and validation of TESS objects. The tool uses a Bayesian frame-
757
+ work that incorporates prior knowledge of the target star, planet
758
+ occurrence rates, and stellar multiplicity to calculate the proba-
759
+ bility that the transit signal is due to a planet transit or another
760
+ astrophysical source. The resulting FPP quantifies the probabil-
761
+ ity that the transit signal is attributed to something other than
762
+ a transiting planet. Using the TESS Sector 30 2-min cadence
763
+ data, combined with the contrast curve obtained by the Zorro
764
+ speckle imaging, we obtain an FPP of 0.094, which lies above
765
+ the threshold for validation. However, the light curve obtained
766
+ by SAINT-EX provides tighter photometric constraints than the
767
+ TESS data. We modified the input to triceratops to instead
768
+ use the SAINT-EX light curve data. The resulting FPP is 0.001
769
+ using the light curve alone, and when also including the contrast
770
+ curve, the FPP is reduced to 3 × 10−11.
771
+
772
+ We also assessed the candidate with vespa, a similar tool
773
+ that compares model light curves from distributions of eclips-
774
+ ing binaries (EBs), including hierarchical eclipsing binaries
775
+ (HEBs) and background eclipsing binaries (BEBs), plus plan-
776
+ etary transits. The stellar populations first are generated using
777
+ isochrones (Morton 2015a), with inputs of the star’s parallax,
778
+ broadband magnitudes from Table 2, and stellar effective tem-
779
+ perature and metallicity given in Table 3. For inputs into vespa,
780
+ we used the transit properties from our MCMC analysis (see
781
+ Sect. 6), the star’s coordinates and a secondary eclipse thresh-
782
+ old of 4.5 ppt to constrain the false positive scenarios. We first
783
+ evaluated using the 2-min cadence TESS Sector 30 data and an
784
+ aperture size of 60(cid:48)(cid:48)(3 TESS pixels), finding an FPP of 0.078,
785
+ again above the threshold required. We then used the higher-
786
+ precision SAINT-EX light curve and an aperture size of 12(cid:48)(cid:48), and
787
+ found the FPP to be less than 10−6.
788
+
789
+ With the triceratops and vespa computed false positive
790
+ probabilities both much less than one per cent, we consider the
791
+ candidate signal to be a validated exoplanet, TOI-2406 b.
792
+
793
+ 7 https://skyview.gsfc.nasa.gov/
794
+
795
+ TESS photometry
796
+
797
+ In this section we aim to search for additional planets in the
798
+ available data and establish detection limits. To search for extra
799
+ planets, we used our custom pipeline SHERLOCK (Pozuelos et al.
800
+ 2020; Demory et al. 2020)8. The SHERLOCK package provides
801
+ the user with easy access to Kepler, K2, and TESS data by
802
+ searching for and downloading the PDC-SAP flux data from
803
+ NASA’s Mikulski Archive for Space Telescope (MAST). Alter-
804
+ natively, the user may instead provide the data in a .csv file if
805
+ needed. Then, due to the associated risk of removing transit sig-
806
+ nals, in particular short and shallow ones, SHERLOCK applies a
807
+ multi-detrend approach to the nominal light curve by means of
808
+ the w¯otan package (Hippke et al. 2019). Hence, the nominal light
809
+ curve is detrended a number of times using a bi-weight filter by
810
+ varying the window size. This strategy allows the user to max-
811
+ imise the signal detection efficiency (SDE) and the S/N of the
812
+ transit search, which is performed over the nominal light curve,
813
+ jointly with the new detrended light curves, by means of the
814
+ transit least squares (TLS) package (Hippke & Heller
815
+ 2019). TLS uses an analytical transit model based on the stellar
816
+ parameters and is optimised for the detection of shallow periodic
817
+ transits. Once a transit signal is detected with a minimum S/N of
818
+ 5, SHERLOCK implements a mask for such a candidate and keeps
819
+ searching in a new run. This operation is repeated until no more
820
+ signals with S /N (cid:62) 5 are found in the dataset.
821
+
822
+ For Sector 3 we made use of the FFI light curve described
823
+ in Sect. 2. For Sector 30 we used the 30- and 2-min cadence
824
+ light curves provided directly by SHERLOCK, which utilises the
825
+ ELEANOR package to access the FFI (Feinstein et al. 2019). In
826
+ all cases, we recovered the candidate issued by TESS with an
827
+ orbital period of 3.07 d in the first run. However, no other signals
828
+ were found, suggesting that: (1) no other planets are present in
829
+ the system; (2) if they do exist, they do not transit; or (3) they
830
+ exist and transit, but the photometric precision of the dataset is
831
+ not enough to detect them, or they have longer periods than the
832
+ ones explored with the dataset in hand. If scenario (2) is true,
833
+ extra planets might be detected by radial velocity follow-up, as
834
+ discussed in Sect. 7.
835
+
836
+ To evaluate scenario (3), we studied the detection limits of
837
+ the current dataset by performing injection-and-recovery exper-
838
+ iments over the 2-min cadence PDC-SAP light curve of Sector
839
+
840
+ 8 The SHERLOCK (Searching for Hints of Exoplanets
841
+ from
842
+ Lightcurves of space-based seekers) code is fully available on GitHub:
843
+ https://github.com/franpoz/SHERLOCK
844
+
845
+ A97, page 7 of 17
846
+
847
+ 0h35m15s10s-3°21'22'23'RA (J2000)Dec. (J2000)POSS I Red (1951.6)0h35m15s10sRA (J2000)POSS II Red (1995.9)0h35m15s10sRA (J2000)SDSS i (2008.8)0h35m15s10sRA (J2000)SAINT-EX z0 (2020.9) A&A 653, A97 (2021)
848
+
849
+ emcee (Foreman-Mackey et al. 2013) implementation of the
850
+ affine-invariant ensemble sampler (Goodman & Weare 2010).
851
+
852
+ We fitted eight parameters of the system: the planet-to-star
853
+ radius ratio (Rp/R∗), transit epoch (T0), orbital period (P), impact
854
+ parameter (b), eccentricity (e), longitude of periastron (ω) and
855
+ the stellar radius (R∗) and mass (M∗). The planet’s semi-major
856
+ axis in stellar radii (a/R∗) was computed from the stellar den-
857
+ sity (from R∗, M∗) and orbital period, using Kepler’s third law.
858
+ We also fitted the quadratic limb-darkening coefficients (LDCs),
859
+ u1 and u2, for each of the five photometric bands covered by the
860
+ transit data (TESS, V, i(cid:48), z(cid:48), Exo). Gaussian priors for the LDCs
861
+ were computed using PyLDTk (Parviainen & Aigrain 2015),
862
+ which uses the stellar spectrum model library of Husser et al.
863
+ (2013). We increased the widths of the computed priors by a fac-
864
+ tor of five to account for the model-dependent uncertainties. For
865
+ each dataset, we included the out-of-transit baseline as a free
866
+ parameter with a Gaussian prior centred on unity with a width
867
+ of 0.01. Detrending vectors for airmass and the full width at half
868
+ maximum were also fit for the ground-based data. An extra vec-
869
+ tor for sky background was also included for the EDEN/VATT
870
+ data, because observations started during twilight. Each vector
871
+ was generated by multiplying the telescope-recorded property by
872
+ a scale factor, which was fitted in the MCMC. Each scale factor
873
+ had a normal prior centred on zero, with a width of 0.1.
874
+
875
+ 6.2. Results
876
+
877
+ The MCMC detrended and transit modelled light curves are
878
+ shown in Fig. 8. The posterior distributions for the eight fit-
879
+ ted transit properties can be found in Appendix A. The fitted
880
+ transit properties are given in Table 4 with their 1σ confidence
881
+ levels. Most notably, we find the planet to have a radius of
882
+ 2.94 ± 0.17 R⊕ and an eccentricity of 0.26+0.27
883
+ −0.12, which leads to
884
+ longer transits than compared to a circular orbit. We also pre-
885
+ dict some properties based on the estimated planet mass. We
886
+ compared planetary masses predicted by the mass-radius rela-
887
+ tionships of Chen & Kipping (2017), Ning et al. (2018) and
888
+ Kanodia et al. (2019). We find the mass posteriors of both Ning
889
+ et al. (2018) and Kanodia et al. (2019) are considerably lower
890
+ than for Chen & Kipping (2017), and extend down to unphysi-
891
+ cally low masses. We therefore elect to use the Chen & Kipping
892
+ (2017) mass estimate in this work, but reason it could be some-
893
+ what over-estimated and so could be seen as more of an upper
894
+ limit.
895
+
896
+ 6.3. Eccentricity
897
+
898
+ Our transit model fit heavily favours a non-zero eccentricity (see
899
+ Fig. 9) and a longitude of periastron around 270◦ (cos ω = 0).
900
+ This is due to the prior on the stellar density (through the stellar
901
+ radius and mass) and the photometry exhibiting a longer transit
902
+ duration than expected for a circular orbit. Given the old age of
903
+ the host star, one might expect eccentricity dampening to have
904
+ produced a near-circular orbit by now. We discuss this is detail
905
+ in Sect. 7.2. To test the case of a circular orbit, we proceeded with
906
+ another MCMC analysis, similar to the one described above,
907
+ with the exception of fixing the eccentricity at zero. While a
908
+ good fit was obtained, the values of the fitted stellar radius
909
+ (0.227 ± 0.005 R(cid:12)) and mass (0.156 ± 0.008 M(cid:12)) differ from the
910
+ priors by 2 and 1σ, respectively. As shown in Fig. 9, this corre-
911
+ sponds to a much lower stellar density compared to the eccentric
912
+ case, which is at odds with our inferred stellar properties.
913
+
914
+ To confirm our analysis was reliable, we also used
915
+ EXOFASTv2 (Eastman et al. 2013, 2019) to simultaneously fit (a)
916
+
917
+ Fig. 7. Injection-and-recovery tests performed to check the detectabil-
918
+ ity of extra planets in the system. We explored a total of 663 different
919
+ scenarios and five different phases each for a total of 3315 simulations.
920
+ Then, each pixel evaluated about 20 scenarios, that is, 20 light curves
921
+ with injected planets having different Pplanet, Rplanet, and T0. Larger
922
+ recovery rates are presented in yellow and green colours, while lower
923
+ recovery rates are shown in blue and darker hues. Planets smaller than
924
+ 1.5 R⊕ would remain undetected for the full set of periods explored.
925
+
926
+ 30. We made use of the MATRIX ToolKit9. We explored the
927
+ Rplanet–Pplanet parameter space in the ranges of 0.7–4.0 R⊕ with
928
+ steps of 0.2 R⊕, and 0.5–10 d with steps of 0.25 d, for a total of
929
+ 663 different scenarios. MATRIX ToolKit allows the user to
930
+ choose a multi-phase approach (i.e. each scenario is explored
931
+ using a number of different values of T0). For simplicity, we
932
+ assumed the impact parameters and eccentricities of the injected
933
+ planets were zero. It is worth noting that before to start the search
934
+ for new candidates we masked the transit times corresponding to
935
+ TOI-2406 b and we detrended the light curves using a biweight
936
+ filter with a window-size of 0.18 d, which was found to be the
937
+ optimal value during the SHERLOCK’s runs. Moreover, as we
938
+ injected the synthetic signals in the PDC-SAP light curve, the
939
+ signals were not affected by the PDC-MAP systematic correc-
940
+ tions; therefore, the detection limits should be considered as the
941
+ most optimistic scenario (see e.g. Pozuelos et al. 2020; Eisner
942
+ et al. 2020).
943
+
944
+ In our case, we explored five phases for a total of 3315 dif-
945
+ ferent scenarios. A synthetic planet was defined as ‘recovered’
946
+ when its epoch was detected with 1 h accuracy and if its period
947
+ was detected better than 5%. The results are shown in Fig. 7,
948
+ which allowed us to reach several conclusions from this test.
949
+ First, we could hardly detect the presence of planets of any size in
950
+ the range explored with orbital periods (cid:38)4 d, where we obtained
951
+ a recovery rate of (cid:46)20%. Second, for orbital periods ≤4 d we
952
+ could detect smaller planets, down to a minimum size of ∼1.5 R⊕
953
+ for orbital periods ≤1 d. Hence, we may rule out the presence of
954
+ such planets; if they existed and transited, they would be easy
955
+ to detect, with recovery rates ranging from 80 to 100%. Finally,
956
+ planets smaller than 1.5 R⊕ would remain undetected for the full
957
+ set of periods explored.
958
+
959
+ 6. Transit analysis
960
+
961
+ 6.1. MCMC procedure
962
+
963
+ All the light curve data detailed in Sect. 2 were fit simulta-
964
+ neously via an MCMC procedure. To do this, we utilised the
965
+ PyTransit (Parviainen 2015) implementation of the Mandel
966
+ & Agol (2002) quadratic limb-darkening transit model, and the
967
+
968
+ 9 The MATRIX ToolKit (Multi-phase Transits Recovery from
969
+ Injected exoplanets Toolkit) code is fully available on GitHub: https:
970
+ //github.com/martindevora/tkmatrix
971
+
972
+ A97, page 8 of 17
973
+
974
+ 1.02.03.04.05.06.07.08.09.010.0Injected period (days)1.01.52.02.53.03.5Injected radius (R)020406080100Recovery rate (%) R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
975
+
976
+ Fig. 8. Phase-folded transit light curves of TOI-2406 b. From left to right and from top to bottom: TESS Sector 30 2-min-cadence, SAINT-EX z(cid:48),
977
+ LCO i(cid:48), EDEN/VATT V, TRAPPIST-South Exo, and TESS 30-min FFI. Each light curve has been corrected for the median trend and baseline from
978
+ the MCMC fit. The median transit model is also over-plotted. The TESS 2-min data (grey points) have been binned by a factor of 15 to produce the
979
+ black data points, with error bars corresponding to the scatter within each bin.
980
+
981
+ the observed TESS and SAINT-EX light curves with a transit
982
+ model, (b) the broadband photometry and the Gaia EDR3 paral-
983
+ lax (assuming zero AV extinction) with an SED based on MIST
984
+ atmospheres, and (c) constraints on the stellar mass and radius
985
+ with MIST stellar evolution models. Thus, our EXOFAST analy-
986
+ sis allows us to explore the circular and eccentric cases, while
987
+ enforcing the astrophysical constraint of a realistic stellar radius
988
+ for a low-mass M dwarf. We used a Gaussian prior on the metal-
989
+ licity and the parallax, and because TOI-2406 is close to the
990
+ 0.1 M(cid:12) limit of the MIST stellar models, we used a Gaussian
991
+ prior on the mass with the values reported in Sect. 3, which are
992
+ based on stellar models that more appropriate for an M4 dwarf.
993
+ We used uniform priors for the stellar radius and for the effective
994
+ temperature. For the eccentric case, we find very similar results,
995
+ with a non-zero eccentricity preferred (e = 0.35+0.33
996
+ −0.15), while the
997
+
998
+ stellar and planetary and parameters are consistent within their
999
+ 1σ uncertainties with those in Table 4. Forcing the orbit to be cir-
1000
+ cular and, given the stellar radius constraints from the SED and
1001
+ stellar models, EXOFAST is not able to fit the SAINT-EX pho-
1002
+ tometry. In other words, we do not find a solution with a circular
1003
+ orbit that is consistent with a physically realistic radius of the
1004
+ star.
1005
+
1006
+ As detailed in Sect. 3.2, there is good agreement between
1007
+ empirical estimates of the stellar mass and radius and our stellar
1008
+ models. We are therefore confident in the accuracy of our param-
1009
+ eter estimates, and, consequently, also in our stellar density
1010
+ estimate. Thus, given the above, we consider that a circular orbit
1011
+ is not likely for TOI-2406 b. We present our preferred best fit and
1012
+ resulting stellar and planet properties for an eccentric orbit in
1013
+ Table 4. For completeness, we include the circular orbit solution
1014
+
1015
+ A97, page 9 of 17
1016
+
1017
+ 0.960.970.980.991.001.011.02TESS 2-minSAINT-EX0.960.970.980.991.001.011.02LCOEDEN/VATT1.51.00.50.00.51.01.50.960.970.980.991.001.011.02TRAPPIST-S1.51.00.50.00.51.01.5TESS 30-min FFITime from mid-transit (hours)Normalised flux Table 4. Median properties with 1σ confidence levels, from the transit
1018
+ analysis.
1019
+
1020
+ A&A 653, A97 (2021)
1021
+
1022
+ Property
1023
+
1024
+ Fitted parameters:
1025
+ T0 (BJD−2 450 000)
1026
+ P (d)
1027
+ Rp/R∗
1028
+ b (R∗)
1029
+ √
1030
+ √
1031
+
1032
+ e cos ω
1033
+ e sin ω
1034
+ R∗ (R(cid:12))
1035
+ M∗ (M(cid:12))
1036
+
1037
+ Value
1038
+
1039
+ 9115.97547 ± 0.00027
1040
+ 3.0766896 ± 6.5 × 10−6
1041
+ 0.1322 ± 0.0020
1042
+ 0.16+0.15
1043
+ −0.11
1044
+ 0.06+0.45
1045
+ −0.55
1046
+ −0.358+0.111
1047
+ −0.095
1048
+ 0.204 ± 0.011
1049
+ 0.162 ± 0.008
1050
+
1051
+ Limb-darkening:
1052
+ u1 TESS
1053
+ u2 TESS
1054
+ u1 z(cid:48)
1055
+ u2 z(cid:48)
1056
+ u1 i(cid:48)
1057
+ u2 i(cid:48)
1058
+ u1 V
1059
+ u2 V
1060
+ u1 Exo
1061
+ u2 Exo
1062
+
1063
+ Derived parameters:
1064
+ Rp (R⊕)
1065
+ ρ∗
1066
+ a/R∗
1067
+ a (AU)
1068
+ i (◦)
1069
+ e
1070
+ ω (◦)
1071
+ S p (S ⊕)
1072
+ Teq
1073
+
1074
+ (†) (K)
1075
+
1076
+ Predicted parameters:
1077
+ Mp (M⊕)
1078
+ K (m s−1)
1079
+ TSM
1080
+ ESM
1081
+
1082
+ 0.313 ± 0.059
1083
+ 0.39 ± 0.11
1084
+ 0.240 ± 0.045
1085
+ 0.354 ± 0.088
1086
+ 0.337 ± 0.066
1087
+ 0.37 ± 0.11
1088
+ 0.56 ± 0.13
1089
+ 0.21 ± 0.17
1090
+ 0.268 ± 0.064
1091
+ 0.25 ± 0.12
1092
+
1093
+ 2.94+0.17
1094
+ −0.16
1095
+ 26.9+4.9
1096
+ −4.4
1097
+ 24.0+1.0
1098
+ −1.1
1099
+ 0.0228 ± 0.0016
1100
+ 89.63+0.27
1101
+ −0.35
1102
+ 0.26+0.27
1103
+ −0.12
1104
+ 279+47
1105
+ −63
1106
+ 6.55+0.94
1107
+ −0.80
1108
+ 447 ± 15
1109
+
1110
+ 9.1+7.1
1111
+ −4.0
1112
+ 14.9+12.0
1113
+ −6.6
1114
+ 115+87
1115
+ −50
1116
+ 4.12+0.67
1117
+ −0.58
1118
+
1119
+ Notes. (†)Calculated with a Bond albedo of zero.
1120
+
1121
+ in Appendix B, but we emphasise that a circular orbit is not
1122
+ preferred from the current data and analysis, and tends to an
1123
+ unphysical stellar radius.
1124
+
1125
+ 7. Discussion
1126
+
1127
+ 7.1. Formation
1128
+
1129
+ The formation of close-in planets is commonly explained
1130
+ by core accretion and migration (e.g. Alibert et al. 2013;
1131
+ Lambrechts et al. 2019). Both planetesimal-based (Ida & Lin
1132
+ 2004; Mordasini et al. 2012; Emsenhuber et al. 2020b) and
1133
+ pebble-based (Brügger et al. 2018; Bitsch et al. 2019) models are
1134
+
1135
+ A97, page 10 of 17
1136
+
1137
+ Fig. 9. Posterior distributions from the MCMC fitting procedure. Top:
1138
+ Stellar density posterior distributions for both the eccentric and circu-
1139
+ lar orbit cases. The prior distribution is over-plotted (arbitrarily scaled),
1140
+ calculated from the stellar mass and radius in Table 3. Bottom: eccen-
1141
+ tricity posterior distribution for the eccentric orbit case. The cumulative
1142
+ probability curve is over-plotted, showing very little agreement with a
1143
+ circular orbit.
1144
+
1145
+ able to reproduce the observed metallicity correlation for giant
1146
+ planets (Gonzalez 1997; Santos et al. 2004). For lower plane-
1147
+ tary masses, the general metallicity trend weakens (Mayor et al.
1148
+ 2011; Buchhave et al. 2012) but remains positive for planets at
1149
+ P < 10 days (Mulders et al. 2016; Petigura et al. 2018). In the
1150
+ following, we briefly put the detection of a large sub-Neptune
1151
+ around a low-metallicity host star into the perspective of current
1152
+ planet formation models.
1153
+
1154
+ Planetesimal-based models (e.g. Emsenhuber et al. 2020a)
1155
+ assume the bulk of the solid material to coagulate into planetes-
1156
+ imals at an early time. Planetesimal formation models predict a
1157
+ significantly steeper planetesimal surface density due to radial
1158
+ drift of pebbles (Dr ˛a˙zkowska & Alibert 2017; Lenz et al. 2019).
1159
+ This process can increase the amount of solids available in the
1160
+ inner regions of the disk, but the total solid mass in the sys-
1161
+ tem remains correlated to the dust to gas ratio of the disk. As
1162
+ the growth of inward-migrating planets – in this planetesimal
1163
+ accretion scenario – is limited by the solid reservoir interior to
1164
+ its starting location (Ida & Lin 2008), a statistical correlation
1165
+ with the initial dust-to-gas ratio of the disk is predicted by these
1166
+ models. The observed stellar metallicities are taken as a proxy to
1167
+ prescribe the total dust-to-gas ratio of the disk.
1168
+
1169
+ While this is mostly relevant for large planetary masses
1170
+ around solar-type stars, it becomes important for low-mass plan-
1171
+ ets around lower stellar masses. Recent ALMA measurements
1172
+ showed that the disk mass is related to the stellar mass with
1173
+ a steeper-than-linear trend (Pascucci et al. 2016; Ansdell et al.
1174
+ 2017). Therefore, much less material might be present around
1175
+ low-mass stars.
1176
+
1177
+ The recent planetesimal-based simulations of Burn et al.
1178
+ (2021) linearly scale Class I disk-mass estimates (Tychoniec et al.
1179
+ 2018) to stellar masses of 0.1 M(cid:12). The reason for the linear scal-
1180
+ ing is due to an identified evolutionary trend in the Class II
1181
+
1182
+ 203040Stellar density (g/cm3)0200004000060000Bin countEccentric posteriorCircular posteriorPrior0.00.20.40.60.81.0Eccentricity0500010000Bin count0.00.20.40.60.81.0Cumulative probability R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
1183
+
1184
+ measurements (Pascucci et al. 2016; Ansdell et al. 2017) and can
1185
+ therefore be interpolated back in time to get an ‘initial’ relation.
1186
+ The Burn et al. (2021) work is based on the updated Bern model
1187
+ of planet formation (Emsenhuber et al. 2020a) and one of the
1188
+ recent theoretical works addressing low-mass stars and making
1189
+ use of the improved observational disk constraints. The model
1190
+ is well suited for comparison to observations by consistently
1191
+ evolving planetary atmospheres over Gyr timescales in order to
1192
+ calculate planetary radii. The results are in general agreement to
1193
+ studies more focused on individual aspects (e.g. Coleman et al.
1194
+ 2019; Schoonenberg et al. 2019 for the TRAPPIST-1 system, or
1195
+ Miguel et al. 2020 who do not vary the stellar metallicity or
1196
+ dust-to-gas ratio and can therefore not be used for our purposes).
1197
+ In the population synthesis results of Burn et al. (2021) for a
1198
+ 0.1 M(cid:12) star, which are shown in Fig. 10, a strong dependence
1199
+ on the sampled metallicity is present. In these planetesimal-
1200
+ accretion simulations,
1201
+ the conditions around low-metallicity
1202
+ hosts never allows for growth of planets to sizes comparable to
1203
+ TOI-2406 b. Instead, only in a few high-metallicity disks – out
1204
+ of the 1000 cases simulated – some similar-sized planets form.
1205
+
1206
+ Reproducing TOI-2406 b in a planetesimal-based, core-
1207
+ accretion model is very challenging. Given the steeper-than
1208
+ linear ALMA measurements, it is unrealistic to further increase
1209
+ total disk masses. Another pathway to increase growth is a
1210
+ reduction in migration timescales. However, this is only efficient
1211
+ where planets grew to significant mass in the early, in situ growth
1212
+ phase. Given the already accretion-favourable parameter choices
1213
+ for planetesimal sizes and dust opacities in the simulations,
1214
+ we conclude that the detection of TOI-2406 b poses a signifi-
1215
+ cant challenge for core-accretion models based on planetesimal
1216
+ accretion.
1217
+
1218
+ If the main solid accretion channel is dominated by pebbles
1219
+ instead of planetesimals (Ormel & Klahr 2010), growth is limited
1220
+ by the pebble isolation mass, which scales linearly with the stel-
1221
+ lar mass (Liu et al. 2019). Therefore, the limit does not depend
1222
+ (to first order) on metallicity. However, there is an indirect depen-
1223
+ dence in a cooling protoplanetary disk: longer pebble accretion
1224
+ timescales in low-metallicity disks lead to planets reaching the
1225
+ pebble isolation mass at later times. Therefore, the temperature
1226
+ has further decreased, and the scale height is reduced (Bitsch
1227
+ et al. 2019). This is also why the environment of the star can
1228
+ influence the metallicity correlation (Ndugu et al. 2018).
1229
+
1230
+ The analysis in the pebble accretion scenario for low-mass
1231
+ star by Liu et al. (2019) shows some potential to form Earth-
1232
+ mass planets around late M dwarfs in rare cases with only a
1233
+ weak dependence on stellar mass. However, Liu et al. (2019)
1234
+ assumed the presence of millimetre-sized pebbles independent
1235
+ of metallicity. It remains to be explored if grains can coagulate
1236
+ on short-enough timescales to build up the required flux of drift-
1237
+ ing pebbles at low metallicities (Birnstiel et al. 2016) and if seed
1238
+ planetesimals can grow (Johansen et al. 2009, e.g. by stream-
1239
+ ing instability). In principle, dust growth should be significantly
1240
+ slower in low-metallicity environments. However, growth and
1241
+ therefore pebble accretion timescales are also sensitive to tur-
1242
+ bulence, fragmentation thresholds, and disk sizes (Dr ˛a˙zkowska
1243
+ et al. 2021).
1244
+
1245
+ There is a strong tension between planetesimal accretion
1246
+ predictions and TOI-2406 b. In contrast, its presence is more
1247
+ naturally produced in the pebble accretion scenario, which is
1248
+ therefore the favoured mode of accretion for this planet. How-
1249
+ ever, open questions about the required pebble and planetesimal
1250
+ growth remain. Also, we cannot rule out formation via gravita-
1251
+ tional instability, where the planet could form far out in the disk
1252
+ and then migrate and downsize (Nayakshin 2010; Forgan et al.
1253
+
1254
+ Fig. 10. Semi-major axis versus radius distribution of synthetic and
1255
+ observed planets. To put TOI-2406 b into perspective, the popula-
1256
+ tion of predicted core-accretion planets from the 0.1 M(cid:12) case in Burn
1257
+ et al. (2021) is shown. The metallicity of the host stars is colour-
1258
+ coded and used in the theoretical calculations as a proxy for the initial
1259
+ planetesimal-to-gas-mass ratio. The metallicity and planetary radii for
1260
+ TRAPPIST-1 are taken from Delrez et al. (2018b) and Agol et al. (2021),
1261
+ respectively; those for LHS 1140 from Ment et al. (2019) and Lillo-Box
1262
+ et al. (2020); and those for GJ 1214 from Anglada-Escudé et al. (2013).
1263
+ Those systems are in better agreement with the theoretical calculations
1264
+ than TOI-2406 b.
1265
+
1266
+ 2018). TOI-2406 therefore presents a benchmark case for future
1267
+ investigations: a large sub-Neptune planet around a significantly
1268
+ metal-poor M dwarf.
1269
+
1270
+ 7.2. Orbital excitation
1271
+
1272
+ Given the age of the system is expected to be ∼11 Gyr, we would
1273
+ not expect the planet to retain a non-zero eccentricity. Follow-
1274
+ ing Goldreich & Soter (1966), using a tidal quality factor (Q(cid:48))
1275
+ of 105, similar to the value expected for Neptune (0.9–33 × 104,
1276
+ Goldreich & Soter 1966; Banfield & Murray 1992; Zhang &
1277
+ Hamilton 2008), and the predicted planetary mass, we find an
1278
+ eccentricity dampening timescale of 0.25 Gyr. This is much
1279
+ shorter than the expected age of the system, and therefore we
1280
+ would expect the planet’s orbit to have been circularised by tidal
1281
+ dissipation. However, we find an eccentricity of ∼0.2 is needed
1282
+ to fit the transit (see Sect. 6.3). A value for Q(cid:48) of 4.4 × 106 is
1283
+ needed for the timescale to approach the expected age of the
1284
+ system, which is much higher than estimated for Neptune.
1285
+
1286
+ TOI-2406 appears to be the fourth planetary system to host a
1287
+ single (sub-) Neptune on a short (P < 10 d), eccentric orbit; the
1288
+ others being, GJ 436 (Butler et al. 2004), GJ 674 (Bonfils et al.
1289
+ 2007), and TOI-269 (Cointepas et al. 2021). The GJ 436 system
1290
+ has been particularly well studied, which consists of a 0.4 M(cid:12)
1291
+ M dwarf, with a Neptune on an eccentric (e = 0.162 ± 0.004;
1292
+ Lanotte et al. 2014) 2.6 d orbit. For this system, another solution
1293
+ has been proposed where the planet’s eccentricity is caused by
1294
+ interaction with another planet or a bound star (Maness et al.
1295
+ 2007; Ribas et al. 2008; Beust et al. 2012). However, searches
1296
+ for such companions have not been rewarding (Ribas et al. 2009;
1297
+ Ballard et al. 2010; Lanotte et al. 2014). One further possibility
1298
+ is that a recent encounter with a nearby star may have perturbed
1299
+ the orbit, and the circularisation will complete in the next few
1300
+ hundred megayears. Here we highlight a possible trend in these
1301
+
1302
+ A97, page 11 of 17
1303
+
1304
+ 102101100a (au)12345R (R)Synthetic PlanetsTrappist-1LHS 1140 b,cGJ 1214 bTOI-2406 b0.40.20.00.20.4[Fe/H]1101001000Period (days) A&A 653, A97 (2021)
1305
+
1306
+ systems, noting that the host stars are all mid-M dwarfs, pointing
1307
+ to a likely common mechanism allowing eccentricity to be raised
1308
+ and remain high.
1309
+
1310
+ TOI-2406 represents the most extreme difference between
1311
+ age of the system and its tidal circularisation timescale. For
1312
+ TOI-2406 b, it is tempting to explain this disparity by a bound
1313
+ companion to the star. This companion could cause an eccen-
1314
+ tricity excitation via direct dynamical interaction or through
1315
+ Lidov-Kozai cycles (Lidov 1962; Kozai 1962; Wu et al. 2007).
1316
+ An object more massive than an ∼M 8 star is ruled out beyond
1317
+ roughly 6 AU by speckle imaging, and hotter stars within this
1318
+ distance would be seen in the near-infrared spectrum. Therefore,
1319
+ the mass of any potential companion is limited to lower-mass
1320
+ objects, such as a brown dwarf or a second planet in the sys-
1321
+ tem. Such objects could be detected by a future radial velocity
1322
+ programme.
1323
+
1324
+ 7.3. Potential for radial velocity observations
1325
+
1326
+ Deriving the mass of TOI-2406 b would not only give us another
1327
+ validation of the planetary nature but would also allow us to
1328
+ derive the full orbital parameters, such as the argument of perias-
1329
+ tron (ω), as well as better constrain the eccentricity. Better con-
1330
+ straining the eccentricity will in turn help us to put constraints
1331
+ on the dynamical history of this system (see Sect. 7.2).
1332
+
1333
+ The mass we have estimated for TOI-2406 b, combined with
1334
+ the derived orbital parameters, indicates a semi-amplitude of
1335
+ 14.9+12.0
1336
+ −6.6 m s−1. Given the faintness of the star, the detection
1337
+ of this shallow signal is challenging even for state-of-the-art
1338
+ instrumentation and requires a great deal of observing time.
1339
+ For stabilised optical spectrographs like ESPRESSO (Pepe et al.
1340
+ 2010), fed by the 8 m unit telescope of the VLT, we expect a
1341
+ radial-velocity (RV) precision of about 40 m s−1 for a single mea-
1342
+ surement with a 15 min exposure. Thus, 80–100 measurements
1343
+ would be needed for a 3σ detection of the RV signal.
1344
+
1345
+ Stabilised high-resolution infrared spectrographs have
1346
+ become operational in the past years that make it possible to
1347
+ investigate such very red stars more efficiently. For example,
1348
+ using the same telescope aperture and same exposure time as
1349
+ for ESPRESSO, but instead the newly commissioned CRIRES+
1350
+ (Dorn et al. 2014) spectrograph (which will become available
1351
+ in October 2021), would lead to one order of magnitude bet-
1352
+ ter precision. Assuming that the instrumental systematics are
1353
+ well understood (Figueira et al. 2010), a 3σ detection of the RV
1354
+ signal would be possible with only 12 measurements. The com-
1355
+ bined measurements would allow the stellar parameters as well
1356
+ as the v sin i of TOI-2406 to be further constrained. Furthermore,
1357
+ any deviation from the expected RV signal would allow one to
1358
+ draw hints on the presence of possible further bodies orbiting
1359
+ TOI-2406.
1360
+
1361
+ Given this further prospect, the direct mass determination
1362
+ of TOI-2406 b will soon become feasible and allow us to better
1363
+ constrain its composition and atmospheric parameters for future
1364
+ transmission spectroscopy.
1365
+
1366
+ 7.4. Potential for transmission spectroscopy
1367
+
1368
+ One of the primary allures of planets transiting small stars is
1369
+ the potential to study their atmospheres in detail. Relatively
1370
+ cool planets like TOI-2406 b are most amenable to studies via
1371
+ transmission spectroscopy (e.g. Lustig-Yaeger et al. 2019). Thus,
1372
+ we consider here atmospheric follow-up of TOI-2406 b via this
1373
+ technique.
1374
+
1375
+ A97, page 12 of 17
1376
+
1377
+ Fig. 11. Suitability of TOI-2406 b for transmission spectroscopy stud-
1378
+ ies. Top panel: complete sample of known transiting exoplanets. Bottom
1379
+ panel: focuses on the coolest ones amenable to study. TOI-2406 b, high-
1380
+ lighted in both panels, is among the most accessible cool planets for
1381
+ atmospheric studies. Planets are colour-coded by the radius bins of
1382
+ Kempton et al. (2018). Filled circles indicate planets with mass mea-
1383
+ surements, while outlined circles indicate those with masses estimated
1384
+ from empirical mass-radius relations. Some highly accessible planets
1385
+ smaller than Neptune are named and highlighted with black outlines in
1386
+ the bottom panel.
1387
+
1388
+ We assessed the potential for studies of TOI-2406 b in
1389
+ transmission using the transmission spectroscopy metric (TSM;
1390
+ Kempton et al. 2018). We calculated the TSM for all tran-
1391
+ siting exoplanets in the NASA Exoplanet Archive10 with a
1392
+ reported stellar radius and effective temperature, and a plane-
1393
+ tary radius and semi-major axis. Following the definition of the
1394
+ TSM, we estimated planetary masses for those without mea-
1395
+ sured ones using the empirical mass-radius relation (Chen &
1396
+ Kipping 2017; Louie et al. 2018). Using the planet-size bins of
1397
+ Kempton et al. (2018), this translated to samples of 296, 611,
1398
+ 202, and 134 measurements for terrestrials (Rp < 1.5 R⊕), small
1399
+ sub-Neptunes (1.5 < Rp < 2.75 R⊕), large sub-Neptunes (2.75 <
1400
+ Rp < 4.0 R⊕), and sub-Jovians (4.0 < Rp < 10 R⊕), respectively.
1401
+ Using the parameters derived here (Table 4), we also calculated
1402
+ the TSM for TOI-2406 b, a planet that we note falls into the large
1403
+ sub-Neptune size bin at 2σ confidence.
1404
+
1405
+ As seen in Fig. 11, we find that the TSM of TOI-2406 b
1406
+ is higher than those of all but 3.0% of the 1243 planets we
1407
+ surveyed. For the large sub-Neptunes, it is higher than all but
1408
+ 2.0% of the sample. In fact, only four large sub-Neptunes
1409
+ have higher TSMs: GJ 1214 b (Charbonneau et al. 2009), K2-
1410
+ 266 b (Rodriguez et al. 2018), HD 191939 b (Badenas-Agusti
1411
+ et al. 2020), and TOI-1130 b (Huang et al. 2020). These have
1412
+ equilibrium temperatures, assuming zero Bond albedo and full
1413
+ day-night heat redistribution (following the definition of the
1414
+ TSM), of 580–1510 K – significantly hotter than the 447 ± 15 K
1415
+ equilibrium temperature of TOI-2406 b11.
1416
+
1417
+ 10 Accessed 24 Feb. 2021.
1418
+ 11 Using a Bond albedo of 0.3, which is observed for Earth, Uranus, and
1419
+ Neptune and might be more realistic for large sub-Neptunes generally,
1420
+ these quoted equilibrium temperatures decrease by 10%.
1421
+
1422
+ 10002003005002000Equilibrium Temperature (K)102101100101102103Transmission Spectroscopy MetricTOI-2406 b200300400500600Equilibrium Temperature (K)102103Transmission Spectroscopy Metric T1 b T1 c T1 d T1 e T1 f T1 g K2-18 b TOI-1266 c L 98-59 d LHS 1140 b LTT 3780 c GJ 1214 b TOI-178 gTerrestrialsSmall sub-NeptunesLarge sub-Neptunessub-JoviansTOI-2406 b R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
1423
+
1424
+ While the final say in the suitability of TOI-2406 b awaits a
1425
+ mass measurement, it is likely a target uniquely suited for studies
1426
+ of relatively cool atmospheres in the large sub-Neptune regime.
1427
+ Crossfield & Kreidberg (2017) note that the amplitude of atmo-
1428
+ spheric features in transmission for warm Neptune-sized planets
1429
+ correlate with either equilibrium temperatures or bulk H/He
1430
+ mass fractions. They argue that these correlations could point to
1431
+ either hazier atmospheres at lower temperatures or more metal-
1432
+ rich atmospheres for smaller planets (or both). Compared to the
1433
+ warm Neptune sample analysed in that work, TOI-2406 b is both
1434
+ relatively cool and small. Studies of its atmosphere in transmis-
1435
+ sion with HST and JWST could thus be useful for further testing
1436
+ both trends suggested by the current sample of Neptune-sized
1437
+ planets.
1438
+
1439
+ 8. Conclusions
1440
+
1441
+ We have presented the discovery and initial analysis of the planet
1442
+ TOI-2406 b. The system consists of a large sub-Neptune orbiting
1443
+ a low-mass member of the thick disk. It is a challenge to the core
1444
+ accretion model of planet formation, where simulations strug-
1445
+ gle to produce large planets around metal-poor, late-type stars.
1446
+ This is particularly problematic for planetesimal-based mod-
1447
+ els: They may point towards a pebble accretion formation for
1448
+ TOI-2406 b, which is more favourable. The planet also has a sig-
1449
+ nificant non-zero eccentricity at an age far beyond the estimated
1450
+ circularisation timescale. Furthermore, the planet is expected to
1451
+ be a good candidate for transmission spectroscopy in the warm
1452
+ Neptune regime with JWST. However, a stronger prediction of
1453
+ the expected S/N awaits a direct mass detection from radial
1454
+ velocity observations, which are possible with the latest infrared
1455
+ spectrographs at 10-metre-class telescopes.
1456
+
1457
+ Acknowledgements. We thank the anonymous referee for their corrections and
1458
+ help in improving the paper. We warmly thank the entire technical staff of
1459
+ the Observatorio Astronómico Nacional at San Pedro Mártir in México for
1460
+ their unfailing support to SAINT-EX operations, namely: E. Cadena, T. Cal-
1461
+ vario, E. Colorado, B. García, G. Guisa, A. Franco, L. Figueroa, B. Hernández,
1462
+ J. Herrera, E. López, E. Lugo, B. Martínez, J. M. Nuñez, J. L. Ochoa, M. Pereyra,
1463
+ F. Quiroz, T. Verdugo, I. Zavala. B.V.R. thanks the Heising-Simons Foundation
1464
+ for support. Y.G.M.C acknowledges support from UNAM-PAPIIT IG-101321.
1465
+ B.-O. D. acknowledges support from the Swiss National Science Foundation
1466
+ (PP00P2-163967 and PP00P2-190080). R.B. acknowledges the support from
1467
+ the Swiss National Science Foundation under grant P2BEP2_195285. M.N.G.
1468
+ acknowledges support from MIT’s Kavli Institute as a Juan Carlos Torres Fellow.
1469
+ A.H.M.J.T acknowledges funding from the European Research Council (ERC)
1470
+ under the European Union’s Horizon 2020 research and innovation programme
1471
+ (grant agreement n◦ 803193/BEBOP), from the MERAC foundation, and from
1472
+ the Science and Technology Facilities Council (STFC; grant n◦ ST/S00193X/1).
1473
+ T.D. acknowledges support from MIT’s Kavli Institute as a Kavli postdoctoral
1474
+ fellow Part of this work received support from the National Centre for Compe-
1475
+ tence in Research PlanetS, supported by the Swiss National Science Foundation
1476
+ (SNSF). The research leading to these results has received funding from the
1477
+ ARC grant for Concerted Research Actions, financed by the Wallonia-Brussels
1478
+ Federation. TRAPPIST is funded by the Belgian Fund for Scientific Research
1479
+ (Fond National de la Recherche Scientifique, FNRS) under the grant FRFC
1480
+ 2.5.594.09.F, with the participation of the Swiss National Science Fundation
1481
+ (SNF). M.G. and E.J. are F.R.S.-FNRS Senior Research Associate. This pub-
1482
+ lication benefits from the support of the French Community of Belgium in the
1483
+ context of the FRIA Doctoral Grant awarded to MT. We acknowledge the use
1484
+ of public TESS data from pipelines at the TESS Science Office and at the
1485
+ TESS Science Processing Operations Center. We acknowledge the use of pub-
1486
+ lic TESS Alert data from pipelines at the TESS Science Office and at the TESS
1487
+ Science Processing Operations Center. Resources supporting this work were pro-
1488
+ vided by the NASA High-End Computing (HEC) Program through the NASA
1489
+ Advanced Supercomputing (NAS) Division at Ames Research Center for the
1490
+ production of the SPOC data products. Funding for the TESS mission is pro-
1491
+ vided by NASA’s Science Mission Directorate. This research has made use of
1492
+ the Exoplanet Follow-up Observation Program website, which is operated by the
1493
+ California Institute of Technology, under contract with the National Aeronau-
1494
+ tics and Space Administration under the Exoplanet Exploration Program. This
1495
+
1496
+ paper includes data collected by the TESS mission that are publicly available
1497
+ from the Mikulski Archive for Space Telescopes (MAST). We thank the TESS
1498
+ GI program G03274 PI, Ryan Cloutier, for proposing the target of this work
1499
+ for 2-min-cadence observations in Sector 30. This work is based upon obser-
1500
+ vations carried out at the Observatorio Astronómico Nacional on the Sierra
1501
+ de San Pedro Mártir (OAN-SPM), Baja California, México. This work makes
1502
+ use of observations from the LCOGT network. Part of the LCOGT telescope
1503
+ time was granted by NOIRLab through the Mid-Scale Innovations Program
1504
+ (MSIP). MSIP is funded by NSF. This work includes data collected at the Vatican
1505
+ Advanced Technology Telescope (VATT) on Mt. Graham. This paper includes
1506
+ data taken on the EDEN telescope network. We acknowledge support from the
1507
+ Earths in Other Solar Systems Project (EOS) and Alien Earths (grant numbers
1508
+ NNX15AD94G and 80NSSC21K0593), sponsored by NASA. Some of the obser-
1509
+ vations in the paper made use of the High-Resolution Imaging instrument Zorro
1510
+ (Gemini program GS-2020B-LP-105). Zorro was funded by the NASA Exo-
1511
+ planet Exploration Program and built at the NASA Ames Research Center by
1512
+ Steve B. Howell, Nic Scott, Elliott P. Horch, and Emmett Quigley. Zorro was
1513
+ mounted on the Gemini South telescope of the international Gemini Observa-
1514
+ tory, a program of NSF’s OIR Lab, which is managed by the Association of
1515
+ Universities for Research in Astronomy (AURA) under a cooperative agreement
1516
+ with the National Science Foundation. on behalf of the Gemini partnership:
1517
+ the National Science Foundation (United States), National Research Council
1518
+ (Canada), Agencia Nacional de Investigación y Desarrollo (Chile), Ministerio
1519
+ de Ciencia, Tecnología e Innovación (Argentina), Ministério da Ciência, Tec-
1520
+ nologia, Inovações e Comunicações (Brazil), and Korea Astronomy and Space
1521
+ Science Institute (Republic of Korea). This research has made use of the NASA
1522
+ Exoplanet Archive, which is operated by the California Institute of Technol-
1523
+ ogy, under contract with the National Aeronautics and Space Administration
1524
+ under the Exoplanet Exploration Program. This work made use of the following
1525
+ Python packages: astropy (Astropy Collaboration 2013, 2018), lightkurve
1526
+ (Lightkurve Collaboration 2018), matplotlib (Hunter 2007), pandas (Wes
1527
+ McKinney 2010), seaborn (Waskom & The Seaborn Development team 2021),
1528
+ scipy (Virtanen et al. 2020) and numpy (Harris et al. 2020).
1529
+
1530
+ References
1531
+ Agol, E., Dorn, C., Grimm, S. L., et al. 2021, Planet. Sci. J., 2, 1
1532
+ Alam, S., Albareti, F. D., Allende Prieto, C., et al. 2015, ApJS, 219, 12
1533
+ Alibert, Y., Carron, F., Fortier, A., et al. 2013, A&A, 558, A109
1534
+ Allard, F., Homeier, D., & Freytag, B. 2012, Phil. Trans. R. Soc. London Ser. A,
1535
+
1536
+ 370, 2765
1537
+
1538
+ Aller, A., Lillo-Box, J., Jones, D., Miranda, L. F., & Barceló Forteza, S. 2020,
1539
+
1540
+ A&A, 635, A128
1541
+
1542
+ Anglada-Escudé, G., Rojas-Ayala, B., Boss, A. P., Weinberger, A. J., & Lloyd,
1543
+
1544
+ J. P. 2013, A&A, 551, A48
1545
+
1546
+ Ansdell, M., Williams, J. P., Manara, C. F., et al. 2017, AJ, 153, 240
1547
+ Astropy Collaboration (Robitaille, T. P., et al.) 2013, A&A, 558, A33
1548
+ Astropy Collaboration (Price-Whelan, A. M., et al.) 2018, AJ, 156, 123
1549
+ Badenas-Agusti, M., Günther, M. N., Daylan, T., et al. 2020, AJ, 160, 113
1550
+ Bailey, S. 2012, PASP, 124, 1015
1551
+ Ballard, S., Charbonneau, D., Deming, D., et al. 2010, PASP, 122, 1341
1552
+ Banfield, D., & Murray, N. 1992, Icarus, 99, 390
1553
+ Belokurov, V., Penoyre, Z., Oh, S., et al. 2020, MNRAS, 496, 1922
1554
+ Benedict, G. F., Henry, T. J., Franz, O. G., et al. 2016, AJ, 152, 141
1555
+ Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410, 527
1556
+ Bensby, T., Feltzing, S., & Oey, M. S. 2014, A&A, 562, A71
1557
+ Beust, H., Bonfils, X., Montagnier, G., Delfosse, X., & Forveille, T. 2012, A&A,
1558
+
1559
+ 545, A88
1560
+
1561
+ Birnstiel, T., Fang, M., & Johansen, A. 2016, Space Sci. Rev., 205, 41
1562
+ Bitsch, B., Izidoro, A., Johansen, A., et al. 2019, A&A, 623, A88
1563
+ Bonfils, X., Mayor, M., Delfosse, X., et al. 2007, A&A, 474, 293
1564
+ Bonfils, X., Gillon, M., Udry, S., et al. 2012, A&A, 546, A27
1565
+ Boyajian, T. S., von Braun, K., van Belle, G., et al. 2012, ApJ, 757, 112
1566
+ Brasseur, C. E., Phillip, C., Fleming, S. W., Mullally, S. E., & White, R. L. 2019,
1567
+
1568
+ Astrophysics Source Code Library [record ascl:1905.007]
1569
+
1570
+ Brown, T. M. 2001, ApJ, 553, 1006
1571
+ Brown, T. M., Baliber, N., Bianco, F. B., et al. 2013, PASP, 125, 1031
1572
+ Brügger, N., Alibert, Y., Ataiee, S., & Benz, W. 2018, A&A, 619, A174
1573
+ Buchhave, L. A., Latham, D. W., Johansen, A., et al. 2012, Nature, 486, 375
1574
+ Burn, R., Schlecker, M., Mordasini, C., et al. 2021, A&A, in press, https://
1575
+
1576
+ doi.org/10.1051/0004-6361/202140390
1577
+
1578
+ Butler, R. P., Vogt, S. S., Marcy, G. W., et al. 2004, ApJ, 617, 580
1579
+ Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, ArXiv e-prints,
1580
+
1581
+ [arXiv:1612.05560]
1582
+
1583
+ Charbonneau, D., Berta, Z. K., Irwin, J., et al. 2009, Nature, 462, 891
1584
+ Chen, J., & Kipping, D. 2017, ApJ, 834, 17
1585
+ Cointepas, M., Almenara, J. M., Bonfils, X., et al. 2021, A&A, 650, A145
1586
+ Coleman, G. A. L., Leleu, A., Alibert, Y., & Benz, W. 2019, A&A, 631, A7
1587
+
1588
+ A97, page 13 of 17
1589
+
1590
+ A&A 653, A97 (2021)
1591
+
1592
+ Collins, K. A., Kielkopf, J. F., Stassun, K. G., & Hessman, F. V. 2017, AJ, 153,
1593
+
1594
+ 77
1595
+
1596
+ Crossfield, I. J. M., & Kreidberg, L. 2017, AJ, 154, 261
1597
+ Cushing, M. C., Vacca, W. D., & Rayner, J. T. 2004, PASP, 116, 362
1598
+ Cutri, R. M., Wright, E. L., Conrow, T., et al. 2021, VizieR Online Data Catalog:
1599
+
1600
+ II/328
1601
+
1602
+ Delrez, L., Gillon, M., Queloz, D., et al. 2018a, SPIE Conf. Ser., 10700, 107001I
1603
+ Delrez, L., Gillon, M., Triaud, A. H. M. J., et al. 2018b, MNRAS, 475, 3577
1604
+ Demory, B. O., Pozuelos, F. J., Gómez Maqueo Chew, Y., et al. 2020, A&A,
1605
+
1606
+ 642, A49
1607
+
1608
+ Dorn, R. J., Anglada-Escude, G., Baade, D., et al. 2014, The Messenger, 156, 7
1609
+ Dressing, C. D., & Charbonneau, D. 2015, ApJ, 807, 45
1610
+ Dr ˛a˙zkowska, J., & Alibert, Y. 2017, A&A, 608, A92
1611
+ Dr ˛a˙zkowska, J., Stammler, S. M., & Birnstiel, T. 2021, A&A, 647, A15
1612
+ Eastman, J., Gaudi, B. S., & Agol, E. 2013, PASP, 125, 83
1613
+ Eastman, J. D., Rodriguez, J. E., Agol, E., et al. 2019, PASP, submitted
1614
+
1615
+ [arXiv:1907.09480]
1616
+
1617
+ Maness, H. L., Marcy, G. W., Ford, E. B., et al. 2007, PASP, 119, 90
1618
+ Mann, A. W., Brewer, J. M., Gaidos, E., Lépine, S., & Hilton, E. J. 2013, AJ,
1619
+
1620
+ 145, 52
1621
+
1622
+ Mann, A. W., Deacon, N. R., Gaidos, E., et al. 2014, AJ, 147, 160
1623
+ Mann, A. W., Feiden, G. A., Gaidos, E., Boyajian, T., & von Braun, K. 2015,
1624
+
1625
+ ApJ, 804, 64
1626
+
1627
+ Mann, A. W., Dupuy, T., Kraus, A. L., et al. 2019, ApJ, 871, 63
1628
+ Mayor, M., Marmier, M., Lovis, C.,
1629
+
1630
+ et
1631
+
1632
+ al. 2011, ArXiv e-prints
1633
+
1634
+ [arXiv:1109.2497]
1635
+
1636
+ Ment, K., Dittmann, J. A., Astudillo-Defru, N., et al. 2019, AJ, 157, 32
1637
+ Miglio, A., Chiappini, C., Mackereth, J. T., et al. 2021, A&A, 645, A85
1638
+ Miguel, Y., Cridland, A., Ormel, C. W., Fortney, J. J., & Ida, S. 2020, MNRAS,
1639
+
1640
+ 491, 1998
1641
+
1642
+ Minkowski, R. L., & Abell, G. O. 1963, The National Geographic Society-
1643
+
1644
+ Palomar Observatory Sky Survey, ed. K. A. Strand, 481
1645
+
1646
+ Morales, J. C., Mustill, A. J., Ribas, I., et al. 2019, Science, 365, 1441
1647
+ Mordasini, C., Alibert, Y., Benz, W., Klahr, H., & Henning, T. 2012, A&A, 541,
1648
+
1649
+ Eisner, N. L., Barragán, O., Aigrain, S., et al. 2020, MNRAS, 494, 750
1650
+ Emsenhuber, A., Mordasini, C., Burn, R., et al. 2020a, A&A, in press, https:
1651
+
1652
+ /doi.org/10.1051/0004-6361/202038553
1653
+
1654
+ A97
1655
+
1656
+ Morton, T. D. 2012, ApJ, 761, 6
1657
+ Morton, T. D. 2015a, Astrophysics Source Code Library
1658
+
1659
+ Emsenhuber, A., Mordasini, C., Burn, R., et al. 2020b, A&A, submitted
1660
+
1661
+ [record ascl:1503.010]
1662
+
1663
+ [arXiv:2007.05562]
1664
+
1665
+ Morton, T. D. 2015b, Astrophysics Source Code Library
1666
+
1667
+ Feinstein, A. D., Montet, B. T., Foreman-Mackey, D., et al. 2019, PASP, 131,
1668
+
1669
+ [record ascl:1503.011]
1670
+
1671
+ 094502
1672
+
1673
+ Fernandes, C. S., Van Grootel, V., Salmon, S. J. A. J., et al. 2019, ApJ, 879, 94
1674
+ Figueira, P., Pepe, F., Melo, C. H. F., et al. 2010, A&A, 511, A55
1675
+ Foreman-Mackey, D. 2016, J. Open Source Softw., 1, 24
1676
+ Foreman-Mackey, D., Hogg, D. W., Lang, D., & Goodman, J. 2013, PASP, 125,
1677
+
1678
+ 306
1679
+
1680
+ Forgan, D. H., Hall, C., Meru, F., & Rice, W. K. M. 2018, MNRAS, 474, 5036
1681
+ Gaia Collaboration (Brown, A. G. A., et al.) 2018, A&A, 616, A1
1682
+ Gaia Collaboration (Brown, A. G. A., et al.) 2021, A&A, 649, A1
1683
+ Giacalone, S., & Dressing, C. D. 2020, Astrophysics Source Code Library
1684
+
1685
+ [record ascl:2002.004]
1686
+
1687
+ Giacalone, S., Dressing, C. D., Jensen, E. L. N., et al. 2021, AJ, 161, 24
1688
+ Gibbs, A., Bixel, A., Rackham, B. V., et al. 2020, AJ, 159, 169
1689
+ Gillon, M., Jehin, E., Magain, P., et al. 2011, EpJ Web Conf., 11, 06002
1690
+ Ginsburg, A., Sip˝ocz, B. M., Brasseur, C. E., et al. 2019, AJ, 157, 98
1691
+ Goldreich, P., & Soter, S. 1966, Icarus, 5, 375
1692
+ Gonzalez, G. 1997, MNRAS, 285, 403
1693
+ Goodman, J., & Weare, J. 2010, Commun. Appl. Math. Comput. Sci., 5, 65
1694
+ Guerrero, N. M., Seager, S., Huang, C. X., et al. 2021, ApJS, 254, 39
1695
+ Hardegree-Ullman, K. K., Cushing, M. C., Muirhead, P. S., & Christiansen, J. L.
1696
+
1697
+ 2019, AJ, 158, 75
1698
+
1699
+ Harris, C. R., Millman, K. J., van der Walt, S. J., et al. 2020, Nature, 585, 357
1700
+ Hippke, M., & Heller, R. 2019, A&A, 623, A39
1701
+ Hippke, M., David, T. J., Mulders, G. D., & Heller, R. 2019, AJ, 158, 143
1702
+ Howell, S. B., Everett, M. E., Sherry, W., Horch, E., & Ciardi, D. R. 2011, AJ,
1703
+
1704
+ 142, 19
1705
+
1706
+ Mulders, G. D., Pascucci, I., & Apai, D. 2015, ApJ, 798, 112
1707
+ Mulders, G. D., Pascucci, I., Apai, D., Frasca, A., & Molenda- ˙Zakowicz, J. 2016,
1708
+
1709
+ AJ, 152, 187
1710
+
1711
+ Nayakshin, S. 2010, MNRAS, 408, L36
1712
+ Ndugu, N., Bitsch, B., & Jurua, E. 2018, MNRAS, 474, 886
1713
+ Newton, E. R., Charbonneau, D., Irwin, J., et al. 2014, AJ, 147, 20
1714
+ Ning, B., Wolfgang, A., & Ghosh, S. 2018, ApJ, 869, 5
1715
+ Nutzman, P., & Charbonneau, D. 2008, PASP, 120, 317
1716
+ Ormel, C. W., & Klahr, H. H. 2010, A&A, 520, A43
1717
+ Parviainen, H. 2015, MNRAS, 450, 3233
1718
+ Parviainen, H., & Aigrain, S. 2015, MNRAS, 453, 3821
1719
+ Pascucci, I., Testi, L., Herczeg, G. J., et al. 2016, ApJ, 831, 125
1720
+ Pepe, F. A., Cristiani, S., Rebolo Lopez, R., et al. 2010, SPIE Conf. Ser., 7735,
1721
+
1722
+ 77350F
1723
+
1724
+ Petigura, E. A., Marcy, G. W., Winn, J. N., et al. 2018, AJ, 155, 89
1725
+ Pollack, J. B., Hubickyj, O., Bodenheimer, P., et al. 1996, Icarus, 124, 62
1726
+ Pozuelos, F. J., Suárez, J. C., de Elía, G. C., et al. 2020, A&A, 641, A23
1727
+ Rayner, J. T., Toomey, D. W., Onaka, P. M., et al. 2003, PASP, 115, 362
1728
+ Rayner, J. T., Cushing, M. C., & Vacca, W. D. 2009, ApJS, 185, 289
1729
+ Reid, I. N., Brewer, C., Brucato, R. J., et al. 1991, PASP, 103, 661
1730
+ Reiners, A., Zechmeister, M., Caballero, J. A., et al. 2018, A&A, 612, A49
1731
+ Ribas, I., Font-Ribera, A., & Beaulieu, J.-P. 2008, ApJ, 677, L59
1732
+ Ribas, I., Font-Ribera, A., Beaulieu, J.-P., Morales, J. C., & García-Melendo, E.
1733
+
1734
+ 2009, IAU Symp., 253, 149
1735
+
1736
+ Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2015, J. Astron. Teles. Instrum.
1737
+
1738
+ Syst., 1, 014003
1739
+
1740
+ Huang, C. X., Quinn, S. N., Vanderburg, A., et al. 2020, ApJ, 892, L7
1741
+ Hunter, J. D. 2007, Comput. Sci. Eng., 9, 90
1742
+ Husser, T.-O., Wende-von Berg, S., Dreizler, S., et al. 2013, A&A, 553, A6
1743
+ Ida, S., & Lin, D. N. C. 2004, ApJ, 616, 567
1744
+ Ida, S., & Lin, D. N. C. 2008, ApJ, 673, 487
1745
+ Jehin, E., Gillon, M., Queloz, D., et al. 2011, The Messenger, 145, 2
1746
+ Jenkins, J. M. 2002, ApJ, 575, 493
1747
+ Jenkins, J. M., Chandrasekaran, H., McCauliff, S. D., et al. 2010, SPIE Conf.
1748
+
1749
+ Ser., 7740, 77400D
1750
+
1751
+ Rodriguez, J. E., Becker, J. C., Eastman, J. D., et al. 2018, AJ, 156, 245
1752
+ Sabin, L., Gómez Maqueo Chew, Y., Demory, B.-O., Petrucci, R., & Saint-Ex
1753
+ Consortium 2018, in 20th Cambridge Workshop on Cool Stars, Stellar Sys-
1754
+ tems and the Sun, Cambridge Workshop on Cool Stars, Stellar Systems, and
1755
+ the Sun, 59
1756
+
1757
+ Santos, N. C., Israelian, G., & Mayor, M. 2004, A&A, 415, 1153
1758
+ Schoonenberg, D., Liu, B., Ormel, C. W., & Dorn, C. 2019, A&A, 627, A149
1759
+ Seager, S., & Sasselov, D. D. 2000, ApJ, 537, 916
1760
+ Sebastian, D., Pedersen, P. P., Murray, C. A., et al. 2020, SPIE Conf. Ser., 11445,
1761
+
1762
+ Jenkins, J. M., Twicken, J. D., McCauliff, S., et al. 2016, SPIE Conf. Ser., 9913,
1763
+
1764
+ 1144521
1765
+
1766
+ 99133E
1767
+
1768
+ Johansen, A., Youdin, A., & Mac Low, M.-M. 2009, ApJ, 704, L75
1769
+ Kanodia, S., Wolfgang, A., Stefansson, G. K., Ning, B., & Mahadevan, S. 2019,
1770
+
1771
+ ApJ, 882, 38
1772
+
1773
+ Kempton, E. M. R., Bean, J. L., Louie, D. R., et al. 2018, PASP, 130, 114401
1774
+ Kozai, Y. 1962, AJ, 67, 591
1775
+ Lambrechts, M., Morbidelli, A., Jacobson, S. A., et al. 2019, A&A, 627, A83
1776
+ Lanotte, A. A., Gillon, M., Demory, B. O., et al. 2014, A&A, 572, A73
1777
+ Lasker, B. M., Sturch, C. R., McLean, B. J., et al. 1990, AJ, 99, 2019
1778
+ Lenz, C. T., Klahr, H., & Birnstiel, T. 2019, ApJ, 874, 36
1779
+ Li, J., Tenenbaum, P., Twicken, J. D., et al. 2019, PASP, 131, 024506
1780
+ Lidov, M. L. 1962, Planet. Space Sci., 9, 719
1781
+ Lightkurve Collaboration (Cardoso, J. V. d. M., et al.) 2018, Astrophysics Source
1782
+
1783
+ Code Library [record ascl:1812.013]
1784
+
1785
+ Lillo-Box, J., Figueira, P., Leleu, A., et al. 2020, A&A, 642, A121
1786
+ Liu, B., Lambrechts, M., Johansen, A., & Liu, F. 2019, A&A, 632, A7
1787
+ Louie, D. R., Deming, D., Albert, L., et al. 2018, PASP, 130, 044401
1788
+ Lustig-Yaeger, J., Meadows, V. S., & Lincowski, A. P. 2019, AJ, 158, 27
1789
+ Mandel, K., & Agol, E. 2002, ApJ, 580, L171
1790
+
1791
+ Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163
1792
+ Smith, J. C., Stumpe, M. C., Van Cleve, J. E., et al. 2012, PASP, 124, 1000
1793
+ Stassun, K. G., & Torres, G. 2016, AJ, 152, 180
1794
+ Stassun, K. G., & Torres, G. 2021, ApJ, 907, L33
1795
+ Stassun, K. G., Collins, K. A., & Gaudi, B. S. 2017, AJ, 153, 136
1796
+ Stassun, K. G., Corsaro, E., Pepper, J. A., & Gaudi, B. S. 2018, AJ, 155, 22
1797
+ Stassun, K. G., Oelkers, R. J., Paegert, M., et al. 2019, AJ, 158, 138
1798
+ Stumpe, M. C., Smith, J. C., Van Cleve, J. E., et al. 2012, PASP, 124, 985
1799
+ Stumpe, M. C., Smith, J. C., Catanzarite, J. H., et al. 2014, PASP, 126, 100
1800
+ Twicken, J. D., Catanzarite, J. H., Clarke, B. D., et al. 2018, PASP, 130, 064502
1801
+ Tychoniec, Ł., Tobin, J. J., Karska, A., et al. 2018, ApJS, 238, 19
1802
+ Van Grootel, V., Fernandes, C. S., Gillon, M., et al. 2018, ApJ, 853, 30
1803
+ Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nat. Methods, 17, 261
1804
+ Waskom, M., & The Seaborn Development Team 2021, J. Open Source Softw.,
1805
+
1806
+ 6, 3021
1807
+
1808
+ Wes McKinney. 2010, in Proceedings of the 9th Python in Science Conference,
1809
+
1810
+ eds. S. van der Walt & J. Millman, 56
1811
+
1812
+ Wu, Y., Murray, N. W., & Ramsahai, J. M. 2007, ApJ, 670, 820
1813
+ Zhang, K., & Hamilton, D. P. 2008, Icarus, 193, 267
1814
+
1815
+ A97, page 14 of 17
1816
+
1817
+ R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
1818
+
1819
+ 1 Center
1820
+
1821
+ for Space
1822
+
1823
+ and Habitability, University
1824
+
1825
+ of Bern,
1826
+
1827
+ Gesellschaftsstrasse 6, 3012 Bern, Switzerland
1828
+ e-mail: robert.wells@csh.unibe.ch
1829
+
1830
+ 2 Department of Earth, Atmospheric and Planetary Science,
1831
+ Massachusetts Institute of Technology, 77 Massachusetts Avenue,
1832
+ Cambridge, MA 02139, USA
1833
+
1834
+ 3 Kavli Institute for Astrophysics and Space Research, Massachusetts
1835
+
1836
+ Institute of Technology, Cambridge, MA 02139, USA
1837
+
1838
+ 4 Universidad Nacional de Córdoba – Observatorio Astronómico de
1839
+
1840
+ Córdoba, Laprida 854, X5000BGR, Córdoba, Argentina
1841
+
1842
+ 5 Consejo Nacional de Investigaciones Cient´ficas y Técnicas (CON-
1843
+
1844
+ ICET), Buenos Aires, Argentina
1845
+
1846
+ 6 Universidad Nacional Autónoma de México,
1847
+ Astronomía, AP 70-264, CDMX 04510, México
1848
+
1849
+ Instituto de
1850
+
1851
+ 7 Center for Astrophysics and Space Science, University of California
1852
+
1853
+ San Diego, La Jolla, CA 92093, USA
1854
+
1855
+ 8 Max Planck Institute for Astronomy (MPIA), Koenigstuhl 17,
1856
+
1857
+ 69117 Heidelberg, Germany
1858
+
1859
+ 9 Astrobiology Research Unit, Université de Liège, Allée du 6 Août
1860
+
1861
+ 19C, 4000 Liège, Belgium
1862
+
1863
+ 10 Space Sciences, Technologies and Astrophysics Research (STAR)
1864
+ Institute, Universit´de Liège, Allée du 6 Août 19C, 4000 Liège,
1865
+ Belgium
1866
+
1867
+ 11 Department of Physics, Massachusetts Institute of Technology,
1868
+
1869
+ Cambridge, MA 02139, USA
1870
+
1871
+ 12 Universidad Nacional Autónoma de México,
1872
+ Astronomía, AP 106, Ensenada 22800, BC, México
1873
+
1874
+ Instituto de
1875
+
1876
+ 13 Department of Physics & Astronomy, Vanderbilt University, 6301
1877
+
1878
+ Stevenson Center Ln., Nashville, TN 37235, USA
1879
+
1880
+ 14 NASA Ames Research Center, Moffett Field, CA 94035, USA
1881
+ 15 School of Physics & Astronomy, University of Birmingham,
1882
+
1883
+ Edgbaston, Birmimgham B15 2TT, UK
1884
+
1885
+ 16 Steward Observatory, The University of Arizona, 933 N. Cherry
1886
+
1887
+ Avenue, Tucson, AZ 85721, USA
1888
+
1889
+ 17 Lunar and Planetary Laboratory, The University of Arizona, 1639 E.
1890
+
1891
+ University Boulevard, Tucson, AZ 85721, USA
1892
+
1893
+ 18 Oukaimeden Observatory, High Energy Physics and Astrophysics
1894
+
1895
+ Laboratory, Cadi Ayyad University, Marrakech, Morocco
1896
+
1897
+ 19 Graduate Institute of Astronomy, National Central University,
1898
+
1899
+ 300 Jhongda Road, Zhongli, Taoyuan 32001, Taiwan
1900
+
1901
+ 20 Instituto de Astrofísica de Canarias (IAC), Calle Vía Láctea s/n,
1902
+
1903
+ 38200 La Laguna, Tenerife, Spain
1904
+
1905
+ 21 Center for Astrophysics, Harvard & Smithsonian, 60 Garden Street,
1906
+
1907
+ Cambridge, MA, 02138, USA
1908
+
1909
+ 22 Dpto. Física Teórica y del Cosmos, Universidad de Granada, 18071,
1910
+
1911
+ Granada, Spain
1912
+
1913
+ 23 NASA Exoplanet Science Institute, Caltech/IPAC, Mail Code
1914
+ 100-22, 1200 E. California Blvd., Pasadena, CA 91125,
1915
+ USA
1916
+
1917
+ 24 Vatican Observatory Research Group, University of Arizona, 933 N
1918
+
1919
+ Cherry Ave., Tucson AZ, 85721-0065, USA
1920
+
1921
+ 25 Department of Astronomy, Tsinghua University, Beijing 100084,
1922
+
1923
+ PR China
1924
+
1925
+ 26 Department of Astronomy, 501 Campbell Hall, University of
1926
+
1927
+ California at Berkeley, Berkeley, CA, 94720, USA
1928
+
1929
+ 27 Department of Physics & Astronomy, University of California, Los
1930
+
1931
+ Angeles, Los Angeles, CA 90095, USA
1932
+
1933
+ 28 Department of Physics, University of Warwick, Coventry, CV4 7AL,
1934
+
1935
+ UK
1936
+
1937
+ 29 Cavendish Laboratory, JJ Thomson Avenue, Cambridge, CB3 0HE,
1938
+
1939
+ UK
1940
+
1941
+ 30 Space Telescope Science Institute, 3700 San Martin Drive,
1942
+
1943
+ Baltimore, MD, 21218, USA
1944
+
1945
+ 31 Patashnick Voorheesville Observatory, Voorheesville, NY 12186,
1946
+
1947
+ USA
1948
+
1949
+ 32 Department of Aeronautics and Astronautics, MIT, 77 Mas-
1950
+
1951
+ sachusetts Avenue, Cambridge, MA 02139, USA
1952
+ 33 SETI Institute, Mountain View, CA 94043, USA
1953
+ 34 Department of Astrophysical Sciences, Princeton University, 4 Ivy
1954
+
1955
+ Lane, Princeton, NJ 08544, USA
1956
+
1957
+ A97, page 15 of 17
1958
+
1959
+ Appendix A: Transit fit posterior distributions
1960
+
1961
+ A&A 653, A97 (2021)
1962
+
1963
+ Fig. A.1. MCMC posterior distributions for the fitted transit parameters. From left-to-right: the planet-to-star radius ratio, transit epoch
1964
+ (BJD−2,450,000), orbital period (d), impact parameter, eccentricity and longitude of periastron parameterised as
1965
+ e sin ω, and
1966
+ the stellar radius and mass. The radius ratio, epoch, period, and stellar parameters can be seen to be Gaussian-like. However, the impact parameter
1967
+ and the eccentricity and periastron distributions are non-Gaussian due to their degeneracy with one another. The detrending parameters and LDCs,
1968
+ not shown here, are well represented by Gaussian distributions. This figure was made using corner.py (Foreman-Mackey 2016).
1969
+
1970
+ e cos ω and
1971
+
1972
+ √
1973
+
1974
+ √
1975
+
1976
+ A97, page 16 of 17
1977
+
1978
+ 0.00450.00500.00550.00600.0065T09115.970.000660.000680.000700.00072P3.0760.00.20.40.60.8b0.80.40.00.40.8ecos0.750.500.250.000.25esin0.1500.1750.2000.2250.250R*0.1250.1300.1350.140Rp/R*0.140.160.180.20M*0.00450.00500.00550.00600.0065T09115.970.000660.000680.000700.00072P3.0760.00.20.40.60.8b0.80.40.00.40.8ecos0.750.500.250.000.25esin0.1500.1750.2000.2250.250R*0.140.160.180.20M* R. D. Wells et al.: A large sub-Neptune transiting the thick-disk M4 V TOI-2406
1979
+
1980
+ Appendix B: Circular orbit fit properties
1981
+
1982
+ Table B.1. Median properties with 1σ confidence levels, from the transit analysis. We give values for both the eccentric and circular orbit cases.
1983
+ We prefer the use of the eccentric model values in this work. We note that the eccentric orbit values given here are identical to those in Table 4 and
1984
+ are shown as a comparison to the circular orbit values. †Calculated with a Bond albedo of zero.
1985
+
1986
+ Property
1987
+
1988
+ Eccentric orbit
1989
+
1990
+ Circular orbit
1991
+
1992
+ Prior
1993
+
1994
+ Fitted parameters:
1995
+ T0 (BJD−2450000)
1996
+ P (d)
1997
+ Rp/R∗
1998
+ b (R∗)
1999
+ √
2000
+ √
2001
+
2002
+ e cos ω
2003
+ e sin ω
2004
+ R∗ (R(cid:12))
2005
+ M∗ (M(cid:12))
2006
+
2007
+ Limb-darkening:
2008
+ u1 TESS
2009
+ u2 TESS
2010
+ u1 z(cid:48)
2011
+ u2 z(cid:48)
2012
+ u1 i(cid:48)
2013
+ u2 i(cid:48)
2014
+ u1 V
2015
+ u2 V
2016
+ u1 Exo
2017
+ u2 Exo
2018
+
2019
+ Derived parameters:
2020
+ Rp (R⊕)
2021
+ ρ∗
2022
+ a/R∗
2023
+ a (AU)
2024
+ i (◦)
2025
+ e
2026
+ ω (◦)
2027
+ Sp (S⊕)
2028
+ † (K)
2029
+ Teq
2030
+
2031
+ Predicted parameters:
2032
+ Mp (M⊕)
2033
+ K (m/s)
2034
+ TSM
2035
+ ESM
2036
+
2037
+ 9115.97547 ± 0.00027
2038
+ 3.0766896 ± 6.5 × 10−6
2039
+ 0.1322 ± 0.0020
2040
+ 0.16+0.15
2041
+ −0.11
2042
+ 0.06+0.45
2043
+ −0.55
2044
+ −0.358+0.111
2045
+ −0.095
2046
+ 0.204 ± 0.011
2047
+ 0.162 ± 0.008
2048
+
2049
+ 9115.97501 ± 0.00027
2050
+ 3.0766882 ± 6.5 × 10−6
2051
+ 0.1319 ± 0.0018
2052
+ 0.097+0.098
2053
+ −0.068
2054
+ −
2055
+ −
2056
+ 0.227 ± 0.005
2057
+ 0.156 ± 0.008
2058
+
2059
+ N(9115.975, 0.1)
2060
+ N(3.07665, 0.001)
2061
+ U(0.001, 0.4)
2062
+ U(0, 1)
2063
+ U(−1, 1)
2064
+ U(−1, 1)
2065
+ N(0.202, 0.011)
2066
+ N(0.162, 0.008)
2067
+
2068
+ N(0.297, 0.060)
2069
+ N(0.346, 0.115)
2070
+ N(0.223, 0.047)
2071
+ N(0.303, 0.097)
2072
+ N(0.321, 0.071)
2073
+ N(0.377, 0.128)
2074
+ N(0.615, 0.143)
2075
+ N(0.286, 0.182)
2076
+ N(0.330, 0.069)
2077
+ N(0.359, 0.124)
2078
+
2079
+ 0.313 ± 0.059
2080
+ 0.39 ± 0.11
2081
+ 0.240 ± 0.045
2082
+ 0.354 ± 0.088
2083
+ 0.337 ± 0.066
2084
+ 0.37 ± 0.11
2085
+ 0.56 ± 0.13
2086
+ 0.21 ± 0.17
2087
+ 0.268 ± 0.064
2088
+ 0.25 ± 0.12
2089
+
2090
+ 2.94+0.17
2091
+ −0.16
2092
+ 26.9+4.9
2093
+ −4.4
2094
+ 24.0+1.0
2095
+ −1.1
2096
+ 0.0228 ± 0.0016
2097
+ 89.63+0.27
2098
+ −0.35
2099
+ 0.26+0.27
2100
+ −0.12
2101
+ 279+47
2102
+ −63
2103
+ 6.55+0.94
2104
+ −0.80
2105
+ 447 ± 15
2106
+
2107
+ 9.1+7.1
2108
+ −4.0
2109
+ 14.9+12.0
2110
+ −6.6
2111
+ 115+87
2112
+ −50
2113
+ 4.12+0.67
2114
+ −0.58
2115
+
2116
+ 0.315 ± 0.058
2117
+ 0.39 ± 0.11
2118
+ 0.236 ± 0.044
2119
+ 0.337 ± 0.088
2120
+ 0.333 ± 0.066
2121
+ 0.36 ± 0.11
2122
+ 0.54 ± 0.13
2123
+ 0.20 ± 0.17
2124
+ 0.267 ± 0.065
2125
+ 0.27 ± 0.12
2126
+
2127
+ 3.27+0.09
2128
+ −0.08
2129
+ 18.8 ± 0.9
2130
+ 24.0+1.0
2131
+ −1.1
2132
+ 0.0254+0.0012
2133
+ −0.0013
2134
+ 89.77+0.16
2135
+ −0.24
2136
+ 0
2137
+ −
2138
+ 6.56+0.94
2139
+ −0.80
2140
+ 447 ± 15
2141
+
2142
+ 10.8+8.3
2143
+ −4.7
2144
+ 15.8+12.3
2145
+ −6.9
2146
+ 107+83
2147
+ −47
2148
+ 4.10+0.67
2149
+ −0.57
2150
+
2151
+ A97, page 17 of 17
2152
+
2153
+
documents/txt/toi282.txt ADDED
The diff for this file is too large to render. See raw diff
 
documents/txt/toi715.txt ADDED
The diff for this file is too large to render. See raw diff
 
documents/txt/topics.txt ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ The thesis name is 'Application of Machine Learning techniques to the exoplanets research'. Initially it consisted in the
2
+ development of a community pipeline named 'SHERLOCK PIPEline', which allows to search for transiting planetary candidates,
3
+ examine their robustness, validate them, fit the priors and posteriors of the physical system, plan ground-based observations
4
+ and analyze the system's stability. Later on, Machine Learning (ML) procedures and technologies are being applied to
5
+ improve the vetting of selected signals by assessing several metrics goodness (centroids offsets, optical ghosts effects,
6
+ etc).
documents/txt/valerie.txt ADDED
The diff for this file is too large to render. See raw diff
 
documents/txt/watson.txt ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ WATSON (Visual Vetting and Analysis of Transits from Space ObservatioNs) is a lightweight software package that enables a comfortable visual vetting of a transiting signal candidate from Kepler, K2 and TESS missions.
2
+
3
+ Any transiting candidate signal found in a space-based mission could have been potentially generated by different astrophysical sources or even by instrumental artifacts induced into a target's light curve. To rule-out these scenarios, the Science Processing Operation Center (SPOC) of the NASA implemented the Data Validation (DV) Reports, which are showing different metrics to qualify or discard an analyzed candidate. These scenarios are mainly
4
+
5
+ Transit shape model fit
6
+ Odd-even transits checks,
7
+ Weak secondary events
8
+ Centroids shifts
9
+ Optical ghost effects
10
+ Transit source offsets
11
+ Rolling band contamination histograms
12
+ WATSON is also implementing similar but more simplified checks for all of those scenarios (SPOC fits transits models and we just compute the SNR of the possible candidate signal) except the rolling band contamination. In addition, we included a new check comparing the transit SNRs in the different available cadences and also all the single-transit plots computed with the official pipeline aperture and a smaller one. With all of these data, we compute metrics that might alert the scientist about problematic signals not complying with any of the thresholds.
13
+
14
+ For more information check the docs: https://dearwatson.readthedocs.io
15
+
16
+ Code example
17
+ Let's say that you found a new candidate in the TOI-175 with a period of 1.04 days (this is a known false positive from the TESS mission). You could run WATSON easily with the next code chunk:
18
+
19
+ period = 1.0491800670761966
20
+ epoch = 1354.7155898963902
21
+ duration_mins = 54
22
+ depth_ppts = 0.158
23
+ sectors = 'all'
24
+ rp_rstar = 0.01205
25
+ a_rstar = 9
26
+ ra = 124.531756290083
27
+ dec = -68.3129998725044
28
+ # We need to mask the previous candidates to ensure they don't add noise to the analysis
29
+ known_transits_mask = [{'P': 3.6906682312153136, 'T0': 1356.20298505518, 'D': 56},
30
+ {'P': 2.2531273092040185, 'T0': 1354.9020729263193, 'D': 45},
31
+ {'P': 7.450690528437917, 'T0': 1355.2917484673462, 'D': 69}]
32
+ your_execution_dir = os.getcwd()
33
+ object_dir = your_execution_dir + '/results'
34
+ if not os.path.exists(object_dir):
35
+ os.mkdir(object_dir)
36
+ Watson(object_dir).vetting("TIC 307210830", period, epoch, duration_mins, depth_ppts, sectors, rp_rstar,
37
+ a_rstar=a_rstar, cadence=120,
38
+ cpus=os.cpu_count() // 2,
39
+ transits_list=None, ra=ra,
40
+ dec=dec, clean=True, transits_mask=known_transits_mask)
41
+ This same example is done in one of our tutorials. Go and check it out!: TOI-175 tutorial