A hot mini-Neptune in the radius valley orbiting solar analogue HD 110113
H.P. Osborn, D.J. Armstrong, L.D. Nielsen, Karen A. Collins, V. Adibekyan, E. Delgado-Mena, G.W. King, J.F. Otegi, N.C. Santos, S. B. Howell, J. Lillo-Box, C. Ziegler, Coel Hellier, C. Briceño, N. Law, A.W. Mann, N. Scott, G. Ricker, R. Vanderspek, David W. Latham, S. Seager, J.N. Winn, Jon M. Jenkins, Diana Dragomir, Dana R. Louie, Benjamin V. Rackham, Joel Villaseñor, Chris Burke, Tansu Daylan, Ares Osborn, D. Barrado, Dennis M. Conti, Eric L. N. Jensen, S. G. Sousa, S. Hoyer, D. A. Caldwell, Jeffrey C. Smith, David R. Rodriguez, Olivier D. S. Demangeon, Daniel Bayliss, Keivan G. Stassun, Susana C.C. Barros, Edward M. Bryant, D. J. A. Brown, P. Figueira, D.R. Anderson, R. West, F. Bouchy, S. Udry, Peter J. Wheatley, R.F. Díaz, D.L. Pollacco, M. Deleuil, C. Dorn, R. Helled, Paul Strøm
MMNRAS , 1–15 (2021) Preprint 14 January 2021 Compiled using MNRAS L A TEX style file v3.0
A hot mini-Neptune in the radius valley orbiting solar analogueHD 110113
H.P. Osborn ★ , D.J. Armstrong , V. Adibekyan , K.A. Collins ,E. Delgado-Mena , S.B. Howell , C. Hellier , G.W. King , J. Lillo-Box , L.D. Nielsen , J.F. Otegi , N.C. Santos , C. Ziegler ,D.R. Anderson , C. Briceño , C. Burke , D. Bayliss , D. Barrado ,E.M. Bryant , D.J.A. Brown , S.C.C. Barros , F. Bouchy ,D.A. Caldwell , D.M. Conti , R.F. Díaz , D. Dragomir , M. Deleuil ,O.D.S. Demangeon , C. Dorn , T. Daylan , P. Figueira , R. Helled ,S. Hoyer , J.M. Jenkins , E.L.N. Jensen , D.W. Latham , N. Law ,D.R. Louie , A.W. Mann , A. Osborn , D.L. Pollacco , D.R. Rodriguez ,B.V. Rackham , G. Ricker , N.J. Scott , S.G. Sousa , S. Seager ,K.G. Stassun , J.C. Smith , P. Strøm , S. Udry , J. Villaseñor ,R. Vanderspek , R. West , P.J. Wheatley , J.N. Winn The authors’ affiliations are shown in Appendix A.
Accepted 07-Jan-2021
ABSTRACT
We report the discovery of HD 110113 b (TOI-755.01), a transiting mini-Neptune exoplanet ona 2.5-day orbit around the solar-analogue HD 110113 ( 𝑇 eff = 5730K). Using TESS photometryand HARPS radial velocities gathered by the
NCORES program, we find HD 110113 b hasa radius of 2 . ± .
12 R ⊕ and a mass of 4 . ± .
62 M ⊕ . The resulting density of2 . + . − . g cm − is significantly lower than would be expected from a pure-rock world;therefore, HD 110113 b must be a mini-Neptune with a significant volatile atmosphere. Thehigh incident flux places it within the so-called radius valley; however, HD 110113 b was ableto hold onto a substantial (0.1-1%) H-He atmosphere over its ∼ . ± . . ± . ⊕ and a period of 6 . + . − . d. Key words: planets and satellites: detection – stars: individual: HD110113
Since its launch in 2018, NASA’s
TESS mission has attempted to de-tect small transiting planets around bright, nearby stars amenable toconfirmation with radial velocity observations (Ricker et al. 2016).The HARPS spectrograph on the 3.6m telescope at La Silla, Chile(Mayor et al. 2003) has been deeply involved in this follow-up effort,beginning with its first detection, the hot super-Earth Pi Mensae c(Huang et al. 2018), and continuing with the first multi-planet sys-tem (TOI-125 Quinn et al. 2019; Nielsen et al. 2020), ★ E-mail: [email protected]
This unique combination of space-based photometry (whichprovides planetary radius) and precise radial velocities (which pro-vide planetary mass) also allows for the determination of exoplanetdensities, and, therefore, an insight into the internal structure ofworlds outside our solar system. These analyses have revealed adiversity of planet structures in the regime between Earth and Nep-tune, from high-density evaporated giant planet cores like TOI-849b(5.2 g cm − Armstrong et al. 2020), to low-density mini-Neptunessuch as TOI-421 c (Carleo et al. 2020), as well as planets whichfollow a more linear track from rocky super-earths to Neptunesdominated by gaseous envelopes, such as the two inner planets or- © a r X i v : . [ a s t r o - ph . E P ] J a n H.P. Osborn et al. biting 𝜈 Lupi (Kane et al. 2020) and TOI-735 (Cloutier et al. 2020;Nowak et al. 2020).The detection of exoplanets with well-constrained physical pa-rameters can also lead to the discovery of statistical trends withinthe planet population which encode information on planetary for-mation and evolution. The "valley" seen around 1 . ⊕ in Keplerdata (Fulton et al. 2017; Van Eylen et al. 2018) is one such feature.According to current theory planets that first formed with gaseousenvelopes within this valley have, due to heating from either theirstars (e.g. evaporation, Owen & Wu 2017) or from internal sources(e.g. core-powered mass loss, Ginzburg et al. 2018), lost those initialgaseous envelopes, thereby evolving to significantly smaller radii tobecome "evaporated cores". By observing the physical parametersof small, hot exoplanets, the exact mechanisms of this process canbe revealed.In this paper, we present the detection, confirmation and RVcharacterisation of two exoplanets orbiting the star HD 110113— the hot mini-Neptune HD 110113 b and the non-transitingHD 110113 c. The observations from which these planets were de-tected are described in section 2, while the analysis of that data isdescribed in section 3. In section 4 we discuss the validity of theouter planet RV signal (4.1), whether HD 110113 is a solar ana-logue (4.2) the internal structure and evaporation of planet b (4.3& 4.4), and potential future observations of the system (4.5). Wesummarize our conclusions in section 5. HD 110113 was observed during
TESS sector 10 with 2-minute ca-dence for 22.5 days, excluding a 2.5 day gap between
TESS orbitsto downlink data. The lightcurve was extracted using the SPOC(Science Processing Operations Centre; Jenkins et al. 2016) SAP(simple aperture photometry) pipeline. It was then processed us-ing the Pre-Search Data Conditioning (PDC, Stumpe et al. 2012;Smith et al. 2012; Stumpe et al. 2014) pipeline, producing precisedetrended photometry with typical precision of 150 ppm/hr for thisstar, and then searched for exoplanetary candidates with the Tran-siting Planet Search (TPS; Jenkins et al. 2010). This identified astrong candidate with a period of 2.54 d, a depth of only 410 ppmand a Signal to Noise Ratio (SNR) of 7.6. Automated and humanvetting subsequently designated this candidate a planet candidateand it was assigned
TESS
Object of Interest (TOI) 755.01.We inspected the
TESS aperture using tpfplotter (plottedin Figure 1; Aller et al. 2020) to ensure no nearby contaminant starscould be causing the transit. We found five stars within the aperturewith contrast less than 8 mag, with the brightest with a Δ mag ofonly 3.5. However, to cause the observed 410 ppm transit, this starwould need to host eclipses of at least 1%. Furthermore, being morethan 1.2 pix, and therefore almost one full-width-half-maximum(FWHM) of the point-spread function (PSF), away from the targetstar, we would expect to see a significant centroid shift. However,the SPOC data validation modelling (Twicken et al. 2018; Li et al.2019) shows no such shift and suggests the transit occurs within0.25 pixels from the target position . The other stars present are As shown by the SPOC DV report accessed at https://mast . stsci . edu/api/v0 . = mast:TESS/product/tess2019085221934-s0010-s0010-0000000073228647-00212_dvr . pdf . Pixel Column Number P i x e l R o w N u m b e r EN m = -2 m = 0 m = 2 m = 4 m = 6 m = 8 12 345 67 8 910111213 14 151617 18 1920 2122 2324 25 26 TIC 73228647 - Sector 10 F l u x × ( e ) Figure 1.
TESS photometric aperture plotted with tpfplotter (Aller et al.2020). The default TESS aperture used by SAP is overplotted in red, andnearby stars down to Δ mag = also > Δ mag of 6.9–7.9),requiring eclipse depths of 25–75%. Causing the observed transitwith such a blend scenario therefore becomes increasingly unlikelygiven the flat-bottomed transit shape of TOI-755.01. We concludethat a blend scenario from a known contaminant is unlikely, howeverwe pursue additional photometry to confirm. We observed a full transit of TOI-755.01 continuously for 443 min-utes in Pan-STARSS 𝑧 -short band on UTC 2020 March 13 fromthe LCOGT (Brown et al. 2013) 1-m network node at Cerro TololoInter-American Observatory. The 4096 × . (cid:48)(cid:48)
389 per pixel, resulting in a26 (cid:48) × (cid:48) field of view. The images were calibrated by the standardLCOGT BANZAI pipeline (McCully et al. 2018). Photometric datawere extracted using
AstroImageJ (Collins et al. 2017). The meanstellar PSF in the image sequence had a FWHM of 2 . (cid:48)(cid:48)
8. Circu-lar apertures with radius 3 . (cid:48)(cid:48) ∼ (cid:48) from the target star. To account for pos-sible contamination from the wings of neighboring star PSFs, wesearched for NEBs out to 2 . (cid:48) 𝑧 -short band, we searched an extra 0.5 mag-nitudes fainter (down to TESS -band magnitude 18.5).The brightness and distance limits resulted in a search forNEBs in 90 Gaia DR2 stars, which includes all stars marked inred in Figure 1 and a further 67 contaminants with Δ mag > MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 by taking into account both the difference in magnitude relative toTOI-755 and the distance to TOI-755 (to account for the estimatedfraction of the star’s flux that would be contaminating the TOI-755SPOC aperture). If the RMS of the 10-minute binned light curveof a neighboring star is more than a factor of 3 smaller than theexpected NEB depth, we consider an NEB to be tentatively ruledout in the star over the observing window. We then visually inspecteach neighboring star’s light curve to ensure no obvious eclipse-likesignal. The LCOGT data rule out possible contaminating NEBs atthe SPOC pipeline nominal ephemeris and over a -1 . 𝜎 to +2 . 𝜎 ephemeris uncertainty window. By process of elimination, we con-clude that the transit is indeed occurring in TOI-755, or a star soclose to TOI-755 that it was not detected by Gaia DR2, or the eventoccurred outside our observing window. Although detecting the transits of HD 110113 b required precisespace-based photometry, ground-based photometric surveys haveobserved HD 110113 and can provide constraints on stellar vari-ability, and therefore an independent measure of the stellar rotationperiod.WASP-South was a wide-field array of 8 cameras forming theSouthern station of the WASP transit-search survey (Pollacco et al.2006). The field of HD 110113 was observed over 150-night spansin each of 2007 and 2008, and then again over 2011 and 2012,acquiring a total of 30 000 photometric data points. WASP-Southwas at that time equipped with 200-mm, f/1.8 lenses, observingwith a 400–700 nm passband, and with a photometric extractionaperture of 48 arcsecs. There are other stars in the aperture aroundHD 110113, but the brightest has Δ mag = Δ mag >
5. Therefore any rotation signal is likely from HD 110113.We searched the data for rotational modulations using the methodsfrom Maxted et al. (2011).The data from 2011 and 2012 show a modulation at a periodof 21 ± ∼ ± ± High-angular-resolution imaging is needed to search for nearbysources that can contaminate the
TESS photometry, resulting inan underestimated planetary radius, or that can be the source ofastrophysical false positives, such as background eclipsing binaries.Through the TESS Follow-Up Program (TFOP), three such imageswere obtained across two telescopes, with the results shown inFigure 3.
We searched for stellar companions to TOI-755 with speckle imag-ing on the 4.1-m Southern Astrophysical Research (SOAR) tele-scope (Tokovinin 2018) on 14 July 2019 UT, observing in Cousins
Figure 2.
Periodograms of the WASP-South data for TOI-755. The top panelshows data from 2007 & 2008 combined, with a significant 42-d periodicity.The lower panels show data from 2011 & 2012, separately and combined,which show more strongly a periodicity of 21-d. The dotted horizontal linesare the estimated 1%-likelihood false-alarm levels.
I-band, a similar visible bandpass as
TESS . More details of the ob-servation are available in Ziegler et al. (2020). The 5 𝜎 detectionsensitivity and speckle auto-correlation functions from the obser-vations are shown in Figure 3. No nearby stars with magnitudesbrighter than 𝐼 =
16 were detected within 3 (cid:48)(cid:48) of HD 110113 in theSOAR observations.
High-resolution speckle interferometric images of HD 110113 wereobtained on 14 January 2020 UT using the Zorro instrumentmounted on the 8-meter Gemini South telescope located on thesummit of Cerro Pachon in Chile. Zorro simultaneously observes intwo bands, (832 nm & 562 nm with widths of 40 & 54 nm respec-tively), obtaining diffraction-limited images with inner working an-gles 0.017 and 0.028 arcsec, respectively. The observation consistedof 3-minute sets of 1000 × 𝜎 contrast curves in MNRAS000
High-resolution speckle interferometric images of HD 110113 wereobtained on 14 January 2020 UT using the Zorro instrumentmounted on the 8-meter Gemini South telescope located on thesummit of Cerro Pachon in Chile. Zorro simultaneously observes intwo bands, (832 nm & 562 nm with widths of 40 & 54 nm respec-tively), obtaining diffraction-limited images with inner working an-gles 0.017 and 0.028 arcsec, respectively. The observation consistedof 3-minute sets of 1000 × 𝜎 contrast curves in MNRAS000 , 1–15 (2021)
H.P. Osborn et al. . . . . . . . Angular separation [arcsec] R e l a t i v ec o n t r a s t [ m a g ] SOARZorro 562nmZorro 832nm-0.5 0.0 +0.5-0.5 0.0 +0.5Angular separation [arcsec]-0.5 0.0 +0.5
Figure 3.
Contrast curves and images from Gemini/Zorro (blue & red for562 and 832nm respectively), and SOAR (green). both filters for the Zorro observation and includes an inset show-ing the 832 nm reconstructed image. The resulting contrast limitsreveal that HD 110113 is a single star to contrast limits of 5 to 8magnitudes, ruling out most main sequence companions to the starwithin the spatial limits of ∼
11 to 320 au (for 𝑑 = . Over the course of two observing seasons in 2018 and 2019, a totalof 114 high-resolution spectra were taken with the High AccuracyRadial velocity Planet Searcher (HARPS, Pepe et al. 2002; Mayoret al. 2003) on the ESO 3.4m telescope at La Silla, Chile. Thesespectra were taken as part of the
NCORES program (PI:Armstrong,1102.C-0249) designed to specifically study the internal structureof hot worlds.We used the high-accuracy mode of HARPS with a 1 (cid:48)(cid:48) sci-ence fibre on the star and a second on-sky fibre monitoring thebackground flux during exposure. The nominal exposure time was1800 seconds, with a few exceptions of slightly longer or shorterintegration, depending on observing conditions and schedule.Spectra and RV information were extracted using the offlineHARPS data reduction pipeline hosted at Geneva Observatory. Weuse a flux template matching a G1 star to correct the continuum-slope in each echelle order. The spectra were cross correlated witha binary G2 mask to derive the cross correlation function (CCF)(Baranne et al. 1996), on which we fit a Gaussian function to obtain
Parameter Value Parameter ValueTOI ID TOI-755 R.A. [ ◦ ] 190 . 𝑎 TIC ID 73228647 𝑏 R.A. [hms] 12:40:08.78 𝑎 HD HD 110113 Dec. [ ◦ ] − . 𝑎 HIP HIP 61820 Dec. [dms] -44:18:43.48 𝑎 Gaia ID 𝑎 𝛿 RA [mas yr − ] − . ± . 𝑎 𝜋 [ mas ] . ± . 𝑎 𝛿 DEC [mas yr − ] − . ± . 𝑎 𝑑 [pc] 106 . ± . 𝑒 𝑅 𝑠 [ 𝑅 (cid:12) ] 0 . ± . 𝑒 B 10 . ± . 𝑐 𝑀 𝑠 [ 𝑀 (cid:12) ] 0 . ± . 𝑒 V 10 . ± . 𝑐 log 𝑔 . ± . 𝑒 Gaia 𝐺 . ± . 𝑎 𝑇 eff [K] 5732 ± 𝑒 TESS mag 9 . ± . 𝑏 [ Fe / H ] . ± . 𝑒 J 8 . ± . 𝑐 𝑣 sin 𝑖 [km s − ] 1 . ± . 𝑒 H 8 . ± . 𝑐 𝑃 rot [d] 20 . ± . 𝑓 K 8 . ± . 𝑐 Age [ Gyr ] ± 𝑔 Table 1.
Stellar parameters. 𝑎 From Gaia DR2(Brown et al. 2018). 𝑏 Fromthe
TESS
Input Catalogue v8 (Stassun et al. 2019). 𝑐 Johnson magni-tudes from APASS (Henden et al. 2015). 𝑑 From 2MASS (Skrutskie et al.2006). 𝑒 Derived from HARPS spectra and archival data - see sect 3.1.1. 𝑓 Determined using the GP fit to activity indicators and RVs as described in3.2.1. 𝑔 Derived from [ Y / Al ] abundance-age relation as described in sect3.1.2. RVs, FWHM and contrast. Additionally, we compute the bisector-span (Queloz et al. 2001) of the CCF and spectral indices tracingchromospheric activity (Gomes da Silva et al. 2011; Boisse et al.2009).We reach a typical SNR per pixel of 75 (order 60, 631nm)in individual spectra, corresponding to an RV error of 1.41m s − .The HARPS spectra and derived RVs were accessed and down-loaded through the DACE portal hosted at the University of Geneva(Buchschacher et al. 2015) under the target name HD 110113 . The star’s effective temperature ( 𝑇 eff ), surface gravity (log 𝑔 ), andmetallicity ( [ Fe / H ] ) were derived using a recent version of theMOOG code (Sneden 1973) and a set of plane-parallel ATLAS9model atmospheres (Kurucz 1993). The analysis was done in LTE.The methodology used is described in detail in Sousa et al. (2011)and Santos et al. (2013a). The full spectroscopic analysis is basedon the Equivalent Widths (EWs) of 233 Fe I and 34 Fe II weaklines by imposing ionization and excitation equilibrium. The line-list used was taken from Sousa et al. (2008). We obtained resultingparameters of 𝑇 eff =5732 ± 𝑔 = . ± .
05 and [ Fe / H ] = . ± .
02. To account for potential systematic uncertainties, weincreased the error bars to 50K and 0.05 dex for 𝑇 eff and log 𝑔 respectively.To constrain the physical stellar parameters of HD 110113given the observed information, we applied three techniques.The first technique was to use the main-sequence calibrationsof Torres et al. (2010) which derive 𝑅 𝑠 and 𝑀 𝑠 using polynomial https://dace . unige . ch/radialVelocities/?pattern = HD110113
MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 λ ( μ m)-12-11-10-9 l og λ F λ ( e r g s - c m - ) Figure 4.
Spectral energy distribution of HD 110113. Red symbols representthe observed photometric measurements, where the horizontal bars representthe effective width of the passband. Blue symbols are the model fluxes fromthe best-fit Kurucz atmosphere model (black). functions of 𝑇 eff , log 𝑔 and [ Fe / H ] , which are built using the ob-served properties of calibration stars. Uncertainties were propagatedusing 10000 Monte Carlo draws and the mass was corrected usingthe calibration of Santos et al. (2013b). This produced a mass andradius of 0 . ± . 𝑀 (cid:12) and 0 . ± . 𝑅 (cid:12) respectively,although Torres et al. (2010) suggest minimum uncertainties of0.06 𝑀 (cid:12) and 0.03 𝑅 (cid:12) respectively.The second was using theoretical isochrones (MIST, Choi et al.2016) as well as observed properties (e.g. colours) to constrain stel-lar parameters, which we performed using isoclassify (Huber2017; Berger et al. 2020). Inputs included the derived spectral prop-erties 𝑇 eff , log 𝑔 and [ Fe / H ] , as well as archival data for HD 110113including APASS
B & V magnitudes (Henden et al. 2015),
Gaia par-allax, Gp, Rp, Bp and luminosity (Brown et al. 2018),
SkyMapper ugriz observations (Onken et al. 2020) and
JHK observa-tions (Skrutskie et al. 2006). This resulted in a mass & radius of1 . + . − . 𝑀 (cid:12) and 1 . ± . 𝑅 (cid:12) respectively. The well-constrained nature of the input measurements mean that we arelimited by the gridsize of the theoretical isochrones, which despitean initial array of more than 3 million points, resulted in only 112samples within all available constraints.As a final independent determination of the basic stellar pa-rameters for HD 110113, we performed an analysis of the broadbandspectral energy distribution (SED) of the star together with the Gaia
DR2 parallax (adjusted by + .
08 mas to account for the systematicoffset reported by Stassun & Torres 2018), in order to determinean empirical measurement of the stellar radius, following the pro-cedures described in Stassun & Torres (2016); Stassun et al. (2017,2018). We pulled the 𝐵 𝑇 𝑉 𝑇 magnitudes from Tycho-2 , the
𝐵𝑉𝑔𝑟𝑖 magnitudes from
APASS , the
𝐽𝐻𝐾 𝑆 magnitudes from , theW1–W4 magnitudes from WISE , the 𝐺𝐺 BP 𝐺 RP magnitudes from Gaia , and the NUV magnitude from
GALEX . Together, the avail-able photometry spans the full stellar SED over the wavelengthrange 0.2–22 𝜇 m (see Figure 4).We performed a fit using Kurucz stellar atmosphere models,with the 𝑇 eff , [Fe/H] and log 𝑔 adopted from the spectroscopic anal-ysis. The only additional free parameter is the extinction ( 𝐴 𝑉 ),which we restricted to the maximum line-of-sight value from thedust maps of Schlegel et al. (1998). The resulting fit is very good (Figure 4) with a reduced 𝜒 of 1.4 and best-fit 𝐴 𝑉 = . ± . 𝐹 bol = . ± . × − erg s − cm − . Taking the 𝐹 bol and 𝑇 eff together with the Gaia
DR2 parallax gives the stellar radius, 𝑅 ★ = . ± . 𝑅 (cid:12) . In addition, we can use the 𝑅 ★ togetherwith the spectroscopic log 𝑔 to obtain an empirical mass estimateof 𝑀 ★ = . ± . 𝑀 (cid:12) .Taken together, all the stellar parameters as derived above areconsistent, and all suggest that HD 110113 is a solar analogue withmass and radius very close to the Sun. As the SED radius measure-ment is least affected by sample size or systematic uncertainty, weassume this as a final radius. Similarly, the mass obtained from thelog 𝑔 and the SED-derived 𝑅 ★ (0 . ± . 𝑀 (cid:12) ) is nearly identical tothat from the MR relationship ( 0 . ± . 𝑀 (cid:12) ), suggesting theyconverge on the same value. We therefore use the mass as definedfrom the offset-corrected Torres et al. (2010) calibrations, with theuncertainty inflated to reflect the typical systematic error (0 . 𝑀 (cid:12) ).To compute the 𝑣 sin 𝑖 from the FWHM, we used the relationsof Dos Santos et al. (2016), who studied the HARPS spectra of alarge number of solar twins. We used this to first estimate the 𝑣 macro from the 𝑇 eff and log 𝑔 derived in section 3.1.1 (3 . ± . − ),and then combined this with the measured FWHM to estimate a 𝑣 sin 𝑖 of 1 . ± .
15 km s − , although the uncertainties here may beunderestimated due to systematic uncertainties. Using the calculated 𝑅 𝑠 , this corresponds to a maximum rotation period ( 𝑃 max ) of 28 ± Stellar abundances of the elements were also derived using the sametools and models as for stellar parameter determination as wellas using the classical curve-of-growth analysis method assuminglocal thermodynamic equilibrium. Although the EWs of the spectrallines were automatically measured with ARES, for the elementswith only two to three lines available we performed careful visualinspection of the EWs measurements. For the derivation of chemicalabundances of refractory elements, we closely followed the methodsdescribed in the literature (e.g. Adibekyan et al. 2012, 2015; DelgadoMena et al. 2014, 2017). Abundances of the volatile elements, Oand C, were derived following the method of Delgado Mena et al.(2010); Bertran de Lis et al. (2015a). Since the two spectral linesof oxygen are usually weak and the 6300.3 Å line is blended withNi and CN lines, the EWs of these lines were manually measuredwith the task splot in IRAF. Lithium and sulfur abundances werederived by performing spectral synthesis with MOOG, following theworks by Delgado Mena et al. (2014) and Costa Silva et al. (2020)respectively. Both abundance indicators are very similar to the solarvalues. All the [X/H] ratios are obtained by doing a differentialanalysis with respect to a high S/N solar (Vesta) spectrum fromHARPS. The stellar parameters and abundances of the elements arepresented in Table 2.We find that the [X/Fe] ratios of most elements are close to solaras expected for a star with this metallicity whereas [O/Fe] and [C/Fe]are slightly subsolar, since these ratios tend to slightly decreaseabove solar metallicity (e.g. Bertran de Lis et al. 2015b; Franchiniet al. 2020). Moreover, we used the chemical abundances of someelements to derive ages through the so-called chemical clocks (i.e.certain chemical abundance ratios which have a strong correlationwith age). We applied the 3D formulas described in Delgado Menaet al. (2019), which also consider the variation in age producedby the effective temperature and iron abundance. The chemicalclocks [Y/Mg], [Y/Zn], [Y/Ti], [Y/Si], and [Y/Al] were derived.
MNRAS000
MNRAS000 , 1–15 (2021)
H.P. Osborn et al.
Parameter Value Error
Abundances
A(Li) 1 .
09 0 . [ Fe / H ] .
14 0 . [ S / H ] .
03 0 . [ Na / H ] .
141 0 . [ Mg / H ] .
129 0 . [ Al / H ] .
105 0 . [ Si / H ] .
097 0 . [ Ca / H ] .
092 0 . [ Ti / H ] .
140 0 . [ Cr / H ] .
156 0 . [ Ni / H ] .
130 0 . [ O / H ] − .
012 0 . [ C / H ] .
032 0 . [ Cu / H ] .
116 0 . [ Zn / H ] .
050 0 . [ Sr / H ] .
170 0 . [ Y / H ] .
170 0 . [ Zr / H ] .
152 0 . [ Ba / H ] .
123 0 . [ Ce / H ] .
120 0 . [ Nd / H ] .
135 0 . Derived Abundance Ratios
Mg/Si 1 .
32 0 . .
08 0 . .
23 0 . Ages [ Y / Mg ] Age [Gyr] 4 .
09 0 . [ Y / Ti ] Age [Gyr] 4 .
09 0 . [ Y / Zn ] Age [Gyr] 3 .
29 0 . [ Y / Si ] Age [Gyr] 3 .
95 0 . [ Y / Al ] Age [Gyr] 4 .
00 0 . Table 2.
Derived stellar abundances. [Y/X] based ages using the3D formula of Delgado Mena et al. (2019) (Table 10: age & 𝑎 + 𝑏 × 𝑇 eff + 𝑐 ×[ Fe / H ]+ 𝑑 ×[ Y / Mg ] ). We selected the [Y/Al] age, 4.0 ± All activity indicators showed clear signs of stellar variability, likelydue to the presence of starspots. To remove this stellar activity, wefirst turned to linear decorrelation of the RV signal using activ-ity indicators. The FWHM and S-index showed the clearest ro-tational signals, so we selected these and used the decorrelationtechnique provided with the DACE spectroscopy Python package(Buchschacher et al. 2015) . Despite this decorrelation removingmuch of the stellar variability signal, the peak at ∼
22d remainedthe single strongest signal in the radial velocity time series (seeFigure 5). To remove the rotation signal at 23 . ± .
08 d, we fitteda 5-parameter Keplerian model (with eccentricity 𝑒 , argument ofperiastron Ω , & semi-amplitude 𝐾 as free parameters, with period 𝑃 and time of transit 𝑡 constrained from the periodogram). Thenext strongest signals were at 6 . ± .
03 d and 2 . ± . https://dace . unige . ch/tutorials/?tutorialId = with amplitudes of 3 . ± .
31 m s − and 2 . ± .
31 m s − respec-tively. This was followed by signals on longer periods, which aremost likely spurious due to rotational and observational aliases.Although this linear decorrelation and Keplerian-fitted rotationperiod was able to reveal the planetary RV signals, stellar variabilitycannot in general be modelled as a Keplerian. Instead we turned to aGaussian process (GP) to model the impact of rotation on the RVs.GPs have frequently been used in the analysis of radial velocitiesaffected by activity (e.g. Haywood et al. 2014; Dumusque et al.2019). One GP kernel well-suited to stellar rotation is a mix ofsimple harmonic oscillator (SHO) terms corresponding to 𝑃 rot and 𝑃 rot /
2, which we built using exoplanet and celerite packages .In order to limit the impact of the GP on the planetary RVsignal, we fitted activity indicators and RV time-series simultane-ously with the same GP kernel, as these should follow the sameunderlying variations with the exception of planetary reflex motion.A similar approach was previously used by Grunblatt et al. (2015)to model stellar variability in the Kepler-78b system, and by SuárezMascareño et al. (2020) to find an outer candidate orbiting ProximaCentauri. By explicitly linking the variation found across activityindicators and RVs, this method has the same effect as "training" aGP on an activity indicator (e.g. Dumusque et al. 2019). However, itavoids having to run multiple models consecutively and transfer theoutput PDF of a training sample into a second model—a processwhich loses information intrinsic to the likely non-Gaussian distri-butions of the GP hyper-parameters as well as information aboutthe correlations between parameters. This technique also enablesthe use of multiple time-series. In this case, we chose S-index andFWHM to co-fit the covariance function with the RVs, as theseshowed the clearest rotation signal.To achieve this, the hyper-parameters for rotation period, mixfactor between 𝑃 rot and 𝑃 rot / 𝑄 ), and thedifference in signal quality between modes ( Δ 𝑄 ) were kept con-stant between S-index, FWHM and RV time-series, while the signalamplitude and mean, which are not shared across parameters, wereset as separate parameters. For each time-series we also used a jit-ter term to model noise not included by measurement errors andto prevent GP over-fitting. All hyper-parameters were given broadpriors, although the rotation period was constrained to the valueobtained from a Lomb-Scargle periodogram (Lomb 1976; Scargle1982) with a standard deviation of 20%. All parameter priors arelisted in Table B1.We also noted that the FWHM errors produced by the HARPSpipeline appeared over-estimated—more than twice the estimatederror derived from the median absolute difference between mea-surements. Therefore the FWHM errors were multiplied by a factorof 0.4386 such that the median error matched the point-to-pointRMS as calculated from the median absolute difference.While we used the GPs to model the covariance between pointsin each timeseries, a mean function is also required to calibrate theaverage value over time, which we applied separately to each ofthe three timeseries. A 2-parameter (i.e. linear) trend term wasincluded to model potential long-term drift in the RVs, althoughthe resulting gradient was not significant ( − . ± .
73 m s − d − ).Single-parameter mean values were included to model the offset ofS-index and FWHM from zero. We used the exoplanet.gp.terms.RotationTerm implementationMNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 P o w e r P=23.6d*
FAP=1\%1-day sampling alias P o w e r P=6.7d10 Period (d)0.00.10.20.30.4 P o w e r P=2.54d
Figure 5.
Periodograms of RVs after linear decorrelation with S-index andFWHM. The upper panel shows the raw periodogram, while subsequentpanels show the periodogram after the removal of the previously markedpeak. The 2.54 d peak is accompanied by a significant peak at the 1-daysampling alias (1.65 d), but the knowledge of a 2.54 d planet in the
TESS photometry breaks this degeneracy. The remaining peaks in the final peri-odogram are likely due to sampling aliases associated with the ∼
60 d spanof observations.
We downloaded the
PDC_SAP lightcurve from the Mikulski Archivefor Space Telescopes (MAST). As high-resolution imaging revealedno close stellar neighbours missed by e.g. the
TESS input cata-logue (Stassun et al. 2019), we made the assumption that the PDC-extracted and dilution-corrected lightcurve for this target was accu-rate. We then normalised the
PDC_SAP timeseries by its medianand masked anomalous flux points from the timeseries by cuttingdata more than 4 . 𝜎 different from both preceding and succeedingneighbours.We initially tried to use the same celerite GP kernel to pre-dict both RV and photometric time-series deviations. This proved tonot be possible, likely because the effect of stellar variability on pho-tometry is not necessarily at the same timescale as for RVs (Aigrainet al. 2012). Similarly, although a Lomb-Scargle periodogram of theraw
TESS lightcurve does show a peak with a period around 25d,the processed
PDC_SAP lightcurve is flat, likely as variability on theorder of a
TESS orbit ( ∼
14 d) is removed during processing.The remaining variability is therefore likely to be the resultof stellar granulation, which is well-suited to be modelled with asingle GP SHO kernel with quality 𝑄 = /√ 𝜔 & 𝑆 ) and priors for the combined analysis and reduce the possibility of the GPs attempting to model the transits themselves,we first fitted this GP to the photometry with planetary transitscut. The interpolated posterior distributions from this analysis thenprovided the priors for the combined analysis. A jitter term wasalso included to model the effect of high-frequency noise not fullyencapsulated by the photon noise (e.g. stellar & spacecraft jitter).We modelled the limb darkening using two approaches: onewhere limb darkening is a free parameter, reparameterised using theapproach of Kipping (2013b) and fitted to the transit with uninfor-mative priors that cover the physical parameter space; and anotherwhere the expected theoretical limb darkening parameters for thestar as generated by Claret (2017) are used as priors for the analysis.We found the resulting distributions to be consistent, and chose touse the second, constrained approach in the final modelling. Thisused a normal prior with the mean, 𝜇 , set from the theoretical param-eter and 𝜎 set as 0.1 which we chose instead of the uncertainty foundwhen propagating the stellar parameters through the Claret (2017)relation, which was likely too contsraining and did not account forsystematic uncertainties. The radius ratio 𝑅 𝑝 / 𝑅 𝑠 was treated usingthe log amplitude to avoid negative values, and b was reparame-terised with 𝑅 𝑝 / 𝑅 𝑠 following the exoplanet implementation ofEspinoza (2018).As ground-based photometry was not precise enough to ob-serve a transit (see Sect. 2.2), we restrict this analysis to only the TESS photometry and HARPS spectroscopy.
We modelled full Keplerian orbits for the two planets, with eccen-tricity priors according to the Kipping (2013a) beta distribution.Monte Carlo sampling, while able to explore the parameterspace around a best-fit solution, does not deal well with exploringunconstrained parameters with multiple local minima. Therefore, inorder to allow our model to explore a single solution, we includednormal priors on period and 𝑡 using the values and uncertaintiesfrom the TOI catalogue in the case of the 2 .
54d planet, and fromthe RV periodogram in the case of the 6 .
7d planet. In all cases, weartificially inflated these uncertainties to make sure the parameterswere not over-constrained by their priors, which is confirmed bynoting that the posterior distributions are, in all cases, narrowerthan the priors.The combined model, built using the exoplanet (Foreman-Mackey et al. 2020) package, was sampled using the No-U TurnSampler (NUTS) in the Hamiltonian Monte Carlo
PyMC back-end(Salvatier et al. 2016) using 5 independent chains with 2000 stepsand an additional 500 steps burn-in. This produced 10000 indepen-dent samples. Model priors and posteriors are displayed in tableB1. The results from the combined model are shown in tables 3 andB1, with the HARPS RV timeseries and best-fit models shown infigure 6, phase-folded RVs and model shown in figure 7, and
TESS photometry and best-fit light curves shown in figure 8.
The periodogram of the activity-corrected radial velocity timeseriesshowed a clear signal at 6 .
75 d, even stronger than that of the planetat 2 .
54 d (Figure 5). No such signal was found by TESS’ automaticTPS; however, there is a chance such a signal may have been missed.
MNRAS000
MNRAS000 , 1–15 (2021)
H.P. Osborn et al. − . . . S i nd e x F W H M [ m s − ] R V [ m s − ] − ∆ R V Figure 6.
S-index, FWHM and RV timeseries of HD110113 for two seasons of HARPS monitoring, with GP models and 2-sigma uncertainty regions overplottedin green. Below the raw RV timeseries is the GP-removed RV timeseries, with the modelled planetary reflex motion and background trend (turquoise). At thevery bottom the full model residuals are shown, with an RMS of only 1.31 m s − —extremely close to the median HARPS measurement uncertainty (1.36 m s − ). − . − . . . . − R V [ m / s ] − . − . . . . − ∆ R V − . − . . . . − R V [ m / s ] − . − . . . . − ∆ R V Figure 7.
Phase-folded RVs (with the best-fit GP model, linear trend, andthe other planetary signal removed) for HD 110113 b (top) and HD 110113 c(bottom), with model-subtracted residuals. Parameter HD 110113 b HD 110113 cEpoch, 𝑡 [BJD-2457000] 1570 . + . − . . ± . 𝑃 [d] 2 . + . − . . + . − . Semi-major Axis, 𝑎 [AU] 0 . ± .
001 0 . + . − . Orbital Eccentricity, 𝑒 . + . − . . + . − . Argument of periastron, Ω − . ± . − . ± . 𝑅 𝑝 / 𝑅 𝑠 ] 0 . ± .
001 —Radius, 𝑅 𝑝 [ 𝑅 ⊕ ] 2 . ± .
12 —Impact Parameter, b 0 . ± .
22 —Transit duration, 𝑡 𝐷 [d] 0 . + . − . —RV semi-amplitude, 𝐾 [m s − ] 2 . ± .
28 3 . ± . 𝑀 𝑝 [ 𝑀 ⊕ ] 4 . ± .
62 10 . ± . ★ Planet Density, 𝜌 𝑝 [gcm − ] 2 . + . − . —Insolation, 𝑆 [kWm − ] 1001 . ± . . ± . 𝑇 𝑝 [K] † . ± . . ± . Table 3.
Derived planet properties. ★ The mass of planet c refers to the 𝑀 𝑝 sin 𝑖 . † Surface temperature assumes a uniform surface and an albedoof 0.2.
A search using the transit least squares algorithm (Hippke& Heller 2019) on the HD 110113 b-subtracted lightcurve found nosignal around 6.7 d, and a visual inspection of the lightcurve aroundthe likely epochs of transits (given the limits from the RV detection)reveals no candidate dips associated with an outer candidate. Indeed,when running a combined model of two transiting planets, withconstraints on orbits from the RVs, the posteriors for the radius of theouter planet were < .
64 R ⊕ at 1 − 𝜎 which, given the 10 . ± . ⊕ mass of HD 110113 c, would be physically impossible, even with aniron-core. Therefore, we come to the conclusion that HD 110113 cis likely non-transiting.In order to assess whether the RV signal alone warrants callingHD 110113 c a confirmed planet or merely a candidate, we ran MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 − r e l a t i v e flu x [ pp t ] − − − r e l a t i v e flu x [ pp t ] − . − . . . . − Figure 8.
TESS photometry, where black dots represent individual 2-minute cadence data and dark circles (with errorbars) represent 30-minute bins. Upperleft:
TESS
PDC_SAP time series with best-fit GP model (both offset by 3.5ppt), and GP-subtracted lightcurve with the best-fit transit model over-plotted (nooffset). Lower left: residuals, with both GP model and transit models subtracted from the lightcurve. Upper right: phase-folded lightcurve of HD 110113 bzoomed to the transit. Lower right: phase-folded residuals. two combined models with identical priors and with one modelincluding a non-transiting planet around 6 . find_MAP function in PyMC3 to find the maximum likelihood for each model, allowing usto compare the difference in Bayesian Information Criterion ( Δ BIC)between the models. The resulting value of Δ BIC = .
32 clearlyfavours a two-planet model over a single planet model, with Δ BIC >
10 suggesting "Very Strong" evidence over the null hypothesis.Another test for the RV signal of HD 110113 c is the coherenceof the signal over time, as radial velocity variation due to, e.g.,stellar variability is not likely to remain coherent over multipleobserving seasons. We verified this two ways using the decorrelatedand rotation-subtracted radial velocities previously used to formRV periodograms (see Figure 5). First we processed each seasonindividually, finding that the signals at 2.451 d and 6.75 d coincidewith peaks during both seasons, albeit at lower signal strength. Nextwe applied the Bayesian generalised Lomb-Scargle periodogram(BGLS, Mortier et al. 2015) to subsets of our RV time series to testsignal coherence as per the technique of Mortier & Collier Cameron(2017). Figure 9 shows that the signal of HD 110113 c passes thistest - remaining evident even in datasets with only a handful ofdatapoints.It should be noted that the period of HD 110113 c, at6 . + . − . d, is close to the 𝑃 rot / 𝑃 rot / 𝑃 rot / 𝑃 rot / 𝑃 rot /
3, but this occursat 7 .
25 d—significantly separated from the RV peak at 6 . + . − . . While we confirm the presence of this second planet, as given Δ BIC >
10, the amplitude of the signal may be affected by thepresence of a signal at 𝑃 rot /
3, therefore the mass of HD 110113 cshould be treated as uncertain.Multiple lines of evidence point to the signal of HD 110113 c N p o i n t s − − − − − − − l og p o w e r Figure 9.
A 2D BGLS periodogram of HD 110113 radial velocities (afterdecorrelation and subtraction of the strongest rotation signal) performedon increasing numbers of radial velocity points, as proposed by Mortier &Collier Cameron (2017). Periods which maintain signal-strength and peri-odicity as a function of observation number suggest coherent (and thereforeplanetary) signals. The two white vertical bands show our modelled peri-ods of each planet. The vertical bands seen are due to signal aliases in thesecond observing season due to the data gap, which is marked with a blackhorizontal line. being planetary in origin. Future RV measurements should helpfurther disentangle stellar rotation and the signal amplitude, andmay even reveal new candidates in this system.The majority of short-period multi-planet systems are typicallyaligned with mutual inclinations of only a few degrees (Lissaueret al. 2011; Figueira et al. 2012; Winn & Fabrycky 2015). To in-vestigate whether this could also be true for HD 110113, we usedthe derived impact parameter of planet b and the semi-major axisratio of b & c to calculate the expected impact parameter of planet
MNRAS000
MNRAS000 , 1–15 (2021) H.P. Osborn et al. c in a perfectly co-planar scenario ( 𝑏 𝑐 = . ± .
42) and the mini-mum mutual inclination ( Δ 𝑖 = . + . ◦− . ). Therefore, the HD 110113planetary system is still consistent with an aligned planetary system.Throughout this work we quote 𝑀 𝑝 sin 𝑖 for HD 110113 c.However, a clear non-detection of transits can constrain a planet’sinclination, and therefore also reduce the lower limit on a planet’smass. However, in this case, the reduction in minimum mass causedby assuming 𝑏 > . 𝑀 𝑝 sin 𝑖 . It is also worth noting that planets b & c have an orbitalperiod ratio near 8/3, although harmonics beyond 2 : 1 are highlyunlikely to create measurable TTVs (Deck & Agol 2015). It is remarkable to note just how sun-like HD 110113 is, with aradius, 𝑇 eff and log 𝑔 all within 1-sigma uncertainties of solar values,with the exception of its slightly higher metallicity ( [ Fe / H ] = . ± . ∼
22 d from archival photometry, spectroscopy timeseries,& 𝑣 sin 𝑖 ). Indeed, this rotation rate is marginally faster than theSun (25–26.5 d when measured with HARPS-N and converted tosidereal period, Milbourne et al. 2019). This could be explained bythe fact that HD 110113 is slightly younger, the Sun rotates slowerthan average (Robles et al. 2008), or the presence of short-periodplanets has tidally inhibited the slow-down of HD 110113, althoughthe effect for such small planets is likely to be small (Bolmont et al.2012).Thanks to their similarities, HD 110113 and its planets couldprove a useful comparison to the Sun and the solar system in thefuture. To explore the composition of HD 110113 b, we performed 4-layerinterior structure modelling, using as inputs the mass and radiusdetermined by our joint modelling of
TESS photometry and HARPSRVs. We followed the method of Otegi et al. (2020), which assumesa pure iron core, a silicate mantle, a non-gaseous water layer, anda H-He atmosphere. In order to quantify the degeneracy betweenthe different interior parameters and produce posterior probabilitydistributions, we use a generalized Bayesian inference analysis witha Nested Sampling scheme (e.g. Buchner et al. 2014). The interiorparameters that are inferred include the masses of the pure-ironcore, silicate mantle, water layer and H-He atmosphere. The ratiosof Fe/Si and Mg/Si found in stars is expected to be mirrored in theprotoplanetary material, and therefore in the internal structures ofexoplanets (Dorn et al. 2015). Hence, we use the values found by ourstellar abundance analysis as a proxy for the core-to-mantle ratio.Given the observed molar ratio of Fe/Si (1 . ± .
07, Table 2) ishigher than that of the Sun (0.85, Lodders et al. 2009), we wouldexpect planetary material around HD 110113 to be more iron-richthan Earth.Table 4 lists the inferred mass fractions of the core, mantle,water-layer and H-He atmosphere from the interior models. Due to
Mass [ M ] R a d i u s [ R ] U N E a r t h - l i k e H O Solar SystemExoplanetsHD 110113 b
Figure 10.
Mass-radius diagram of exoplanets with accurate mass and radiusdetermination (Otegi et al. 2019). Also shown are the mass-radius relationsfor Earth-like and pure water compositions.
Table 4.
Inferred interior structure properties of TOI-755b.Constituent With H-He [%] With H O [%] 4-layer [%] 𝑀 core / 𝑀 total + − + − + − 𝑀 mantle / 𝑀 total + − + − + − 𝑀 water / 𝑀 total — 73 + − + − 𝑀 H − He / 𝑀 total . + . − . — 0 . + . − . the nature of the measurements, interior models cannot distinguishbetween water and H-He as the source of low-density material.Therefore, we ran both a 4-layer model and two 3-layer models,which leave out the H O and H-He envelopes, respectively. In thecase of a H-He envelope, we find that the planet is only ∼ + − % water. Such a high water-to-rock ratio ischallenging from formation point of view. Therefore HD 110113 balmost certainly has a significant gaseous envelope. Stars with super-solar metallicities are also less likely to host water-rich planets dueto a higher C/O ratio (Bitsch & Battistini 2020), making a water-richcomposition even less likely.Figure 10 shows the mass radius relation (M-R relation) forEarth-like and pure water compositions (where the pure water linecorresponds to a surface pressure of 1 bar, and without a water-vapor atmosphere). Also shown are exoplanets with accurate massand radius determinations from Otegi et al. (2020). The position ofHD 110113 b makes it one of the lowest-density worlds found with 𝑀 𝑝 < ⊕ , and among a small class of low-density low-massplanets which includes 𝜋 Men c (Huang et al. 2018) and GJ 9827 b(Niraula et al. 2017).
With an insolation of 1001 . ± . − ( ∼ 𝑆 ⊕ ), it isextremely likely that HD 110113 b has been moulded by strong MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 P l a n e t S i ze [ E a r t h r a d ii] typicaluncert. . . . . . . . R e l a t i v e O cc u rr e n ce Figure 11.
The distribution of Kepler planets by both insolation & planetaryradius plot, with underlying occurrence distributions adapted from (Martinezet al. 2019). HD 110113 b is included as a purple star. The best-fit positionsof radius valleys from both Martinez et al. (2019) (dashed) and Van Eylenet al. (2018) (dotted) are plotted in blue, with conversion from period toinsolation performed using the average stellar parameters in the Keplersamples. Typical uncertainties for both HD 110113 c and for the Keplersample are shown in the top left. stellar radiation in some way. This is further suggested by placingHD 110113 b on the insolation-radius plots of Fulton et al. (2017)and Martinez et al. (2019), which clearly show the "radius val-ley" (see Fig. 11. The negative slope of the valley with insolationmeans that, even with a radius of 2 . ± .
12 R ⊕ , HD 110113 b ispositioned exactly within it.Using both rotation and age, we predict a current X-ray lumi-nosity ( 𝐿 𝑥 / 𝐿 bol ) of between 8 . × − (with Prot; Wright et al.2018) and 2 . × − (with age; Jackson et al. 2012). This impliestotal X-ray luminosities on the order of 3 . × to 2 . × ergand mass-loss rates (assuming an energy-limited regime) of between5 × and 9 × gs − (0.026– 0.05 𝑀 ⊕ Gyr − ). This is compa-rable to both GJ 436 b and Pi Men c under similar assumptions(King et al. 2019). Therefore, while it is currently highly irradiated,HD 110113 b is unlikely to currently be losing large quantities of itsH-He atmosphere to space.However, the integrated sum of mass-loss since the planet’sformation is substantial, as young stars are typically far more activeand far more X-ray luminous. We calculate that, assuming the cur-rent mass and radius, as much as 10% of the planet’s mass may havebeen lost through evaporation. The models of Zeng et al. (2019) sug-gest that a 1000 K planet with >
5% hydrogren and a 5 . 𝑀 ⊕ corewould have been > . ⊕ in radius, suggesting that HD 110113 bpotentially started as an extremely low-density Jupiter-radius worldwhich was quickly stripped. How such a low-mass world came topossess such a large gaseous atmosphere raises more questions.In any case, it is highly likely that HD 110113 b started witha thicker atmosphere of H-He, which, due to both evaporative andcore-powered mass-loss, it lost much of over time. However, thisis typically a runaway process: planets which lose the majority oftheir gas (i.e. those in the radius valley) typically lose it all (Owen& Wu 2017). Therefore the main unanswered question is: howdid HD 110113 b escape becoming a naked core devoid of volatileenvelope? Here we propose two solutions to this:1) HD 110113 b started with a large envelope of H-He, perhapsas much as 10%, which was gradually lost to evaporation and core- powered heating over time. But it had just enough gas to walk thetight-rope between keeping hold of a thick atmosphere and beingcompletely stripped such that, at the point that evaporative forcingstopped, HD 110113 b still had ∼
1% of H-He by mass. The modelsof Rogers & Owen (2020, Figure 4) suggest such a scenario ispossible and may occur for planets that start gas-rich with around4% H-He.2) HD 110113 b did lose almost all of its H-He to evaporationand core-powered mass-loss. The current density is therefore ex-plained by the planet having a large water content (e.g. an icy core),with potential out-gassing of a He-depleted secondary atmospherecontributing to the extended radius. Indeed, our composition calcu-lations include water in only solid & liquid states; therefore a thicksteam (or supercritical; Mousis et al. 2020) H O atmosphere couldreduce the density without requiring >
50% H O.One final solution might be that HD 110113 b and HD 110113 cunderwent late-stage migration to their current positions, therebyavoiding much of the evaporative forcing caused by the X-ray emis-sions of the young star. However, there is no theoretical mechanismin which a low-eccentricity 2-planet system could undergo suchlate-stage migration long after the dispersal of the protoplanetarydisc. Instead, multi-planet systems are capabale of undergoing early-stage migration damped by the protoplanetary gas disc (Cresswell& Nelson 2006; Carrera et al. 2019), and massive single planets arethought capable of undergoing late-stage, high-eccentricity scatter-ing onto shorter orbits (Ford & Rasio 2008; Beaugé & Nesvorný2012). We therefore consider a solution through in-situ processesmore plausible than through migration.These two predictions may be testable with future transmissionspectroscopy observations, e.g. with
JWST (Greene et al. 2016;Beichman et al. 2014).
The low-density nature of this hot mini-Neptune, combined withits bright host star, may enable transmission spectroscopy observa-tions. Such measurement could test the hypotheses noted above, andsearch for a low-molecular weight primary atmosphere dominatedby H-He, or a high molecular weight secondary atmosphere domi-nated by an overabundance of water vapour (Bean et al. 2020). Totest this, we computed the emission and transmission spectroscopymetrics from Kempton et al. (2018).We find that, amongst small planets with 𝑅 𝑝 < 𝑅 ⊕ (Akesonet al. 2013) , HD 110113 b ranks in the top 3% most amenable foremission and the top 5% for transmission spectroscopy with JWST .Although, when compared to one of the most favourable
JWST targets: the low-density mini-Neptune GJ 1214 b, HD 110113 b pro-vides only around 10% the SNR in both transmission & emission— as is expected when comparing with a planet whose transits are36 times deeper.HD 110113 b will be also re-observed by
TESS during Sector37 , and could also be observed by ESA’s CHEOPS telescope (Benzet al. 2020), both of which would improve the radius precisionbelow the currently measured value of 7%, thereby improving ourknowledge of the internal structure of HD 110113 b. https://exoplanetarchive . ipac . caltech . edu/cgi-bin/nstedAPI/nph-nstedAPI?table = exoplanets&select = *&format = csv , accessed 2020-Oct-18 https://heasarc . gsfc . nasa . gov/cgi-bin/tess/webtess/wtv . py?Entry = MNRAS000
TESS during Sector37 , and could also be observed by ESA’s CHEOPS telescope (Benzet al. 2020), both of which would improve the radius precisionbelow the currently measured value of 7%, thereby improving ourknowledge of the internal structure of HD 110113 b. https://exoplanetarchive . ipac . caltech . edu/cgi-bin/nstedAPI/nph-nstedAPI?table = exoplanets&select = *&format = csv , accessed 2020-Oct-18 https://heasarc . gsfc . nasa . gov/cgi-bin/tess/webtess/wtv . py?Entry = MNRAS000 , 1–15 (2021) H.P. Osborn et al.
We have presented the detection and confirmation of HD 110113 b,which was initially spotted as TOI-755.01 in
TESS with an SNRof only 7 . 𝜎 and transit depth of 410 ppm. This marks one of thelowest-SNR signals yet to be confirmed from TESS , and is testamentto the unique ability of
TESS to find planet candidates around brightstars which can be redetected and characterised through independentRV campaigns.High-resolution imaging and ground-based photometry rulesout the presence of nearby companions and potential nearby eclips-ing binaries, thereby limiting the number of false-positives and giv-ing us confidence to follow such a low-SNR signal. Our subsequentHARPS campaign obtained more than 100 HARPS spectra in orderto characterise both HD 110113 b and its bright ( 𝐺 = . 𝑇 eff ,log 𝑔 and age. The RV timeseries also revealed strong activity onHD 110113 with a rotation period of 20 . ± . . + . − . d and6 . + . − . d. The inner signal, from a planet with mass 4 . ± . ⊕ , corresponded to the detected TESS candidate with a radius, asmodelled from the
TESS photometry, of 2 . ± .
12 R ⊕ . The outersignal, from a planet with 𝑀 𝑝 sin 𝑖 of 10 . ± . ⊕ did not corre-spond to any transit events in the TESS lightcurve, and therefore islikely non-transiting. We were able to confirm it as a planet throughBayesian model comparison which showed Δ BIC = .
32 in favourof a 2-planet model.The estimated density of HD 110113 b is 2 . + . − . g cm − —far lower than would be expected from a rocky core. By mod-elling four potential constituents—an iron core, silicate mantle, wa-ter ocean and H-He atmosphere—we were able to rule out a gaslesscomposition for HD 110113 b, suggesting that it has between 0.07and 1.5% H-He by mass. This is surprising given HD 110113 b’sposition in the "radius valley" between gaseous mini-Neptunes androcky super-Earths, and we suggest two possibilities for this unex-pectedly low density: either HD 110113 b has a water-rich core andsecondary atmosphere, or it began with a thick H-He envelope andmanaged to retain a small fraction of it despite significant evapora-tion and/or heating. Follow-up spectroscopy observations with thenext generation of telescopes may reveal the answer, as well as farmore about this interesting system. ACKNOWLEDGEMENTS
We thank Raphaëlle Haywood, Maximillian Günther and Francois Bouchyfor discussion on disentangling RV activity from signals.This paper includes data collected by the
TESS mission. Funding for the
TESS mission is provided by the NASA Explorer Program and NASA’sScience Mission directorate. We acknowledge the use of public
TESS
Alertdata from pipelines at the
TESS
Science Office and at the
TESS
ScienceProcessing Operations Center. This paper includes data collected by the
TESS mission, which are publicly available from the Mikulski Archive forSpace Telescopes (MAST).This research has made use of the Exoplanet Follow-up Observation Pro-gram website, which is operated by the California Institute of Technology,under contract with the National Aeronautics and Space Administrationunder the Exoplanet Exploration Program.Resources supporting this work were provided by the NASA High-EndComputing (HEC) Program through the NASA Advanced Supercomputing (NAS) Division at Ames Research Center for the production of the SPOCdata products.This study is based on observations collected at the European SouthernObservatory under ESO programme 1102.C-0249.We thank the Swiss National Science Foundation (SNSF) and the GenevaUniversity for their continuous support to our planet search programs. Thiswork has been in particular carried out in the frame of the National Centrefor Competence in Research
PlanetS supported by the Swiss NationalScience Foundation (SNSF).This publication makes use of The Data & Analysis Center for Exoplanets(DACE), which is a facility based at the University of Geneva (CH)dedicated to extrasolar planets data visualisation, exchange and analysis.DACE is a platform of the Swiss National Centre of Competence inResearch (NCCR) PlanetS, federating the Swiss expertise in Exoplanetresearch. The DACE platform is available at https://dace . unige . ch .This work makes use of observations from the LCOGT network.(Some of the) Observations in the paper made use of the High-ResolutionImaging instrument(s) ‘Alopeke (and/or Zorro). ‘Alopeke (and/or Zorro)was funded by the NASA Exoplanet Exploration Program and built at theNASA Ames Research Center by Steve B. Howell, Nic Scott, Elliott P.Horch, and Emmett Quigley. Data were reduced using a software pipelineoriginally written by Elliott Horch and Mark Everett. ‘Alopeke (and/orZorro) was mounted on the Gemini North (and/or South) telescope ofthe international Gemini Observatory, a program of NSF’s OIR Lab,which is managed by the Association of Universities for Research inAstronomy (AURA) under a cooperative agreement with the NationalScience Foundation. on behalf of the Gemini partnership: the NationalScience Foundation (United States), National Research Council (Canada),Agencia Nacional de Investigación y Desarrollo (Chile), Ministerio deCiencia, Tecnología e Innovación (Argentina), Ministério da Ciência,Tecnologia, Inovações e Comunicações (Brazil), and Korea Astronomy andSpace Science Institute (Republic of Korea).Based in part on observations obtained at the Southern AstrophysicalResearch (SOAR) telescope, which is a joint project of the Ministério daCiência, Tecnologia e Inovações (MCTI/LNA) do Brasil, the US NationalScience Foundation’s NOIRLab, the University of North Carolina at ChapelHill (UNC), and Michigan State University (MSU), and the internationalGemini Observatory, a program of NSF’s NOIRLab, which is managed bythe Association of Universities for Research in Astronomy (AURA) undera cooperative agreement with the National Science Foundation. on behalfof the Gemini Observatory partnership: the National Science Foundation(United States), National Research Council (Canada), Agencia Nacionalde Investigación y Desarrollo (Chile), Ministerio de Ciencia, Tecnologíae Innovación (Argentina), Ministério da Ciência, Tecnologia, Inovações eComunicações (Brazil), and Korea Astronomy and Space Science Institute(Republic of Korea).H.P.O. acknowledges support from NCCR/Planet-S via the CHESSfellowship.D.J.A. acknowledges support from the STFC via an Ernest RutherfordFellowship (ST/R00384X/1).V.A., E.D.M., N.C.S., O.D.S.D. & S.C.C.B. acknowledge support by FCT -Fundação para a Ciência e a Tecnologia (Portugal) through national fundsand by FEDER through COMPETE2020 - Programa Operacional Com-petitividade e Internacionalização by these grants: UID/FIS/04434/2019;UIDB/04434/2020; UIDP/04434/2020; PTDC/FIS-AST/32113/2017 &POCI-01-0145-FEDER-032113; PTDC/FIS-AST/28953/2017 & POCI-01-0145-FEDER-028953.V.A. and E.D.M. further acknowledge the support from FCT throughInvestigador FCT contracts IF/00650/2015/CP1273/CT0001 andIF/00849/2015/CP1273/CT0003.O.D.S.D. and S.C.C.B. are supported through Investigador contract (DL57/2016/CP1364/CT0004) funded by FCT.J.L-B. is supported by the Spanish State Research Agency (AEI) ProjectsNo.ESP2017-87676-C5-1-R and No. MDM-2017-0737 Unidad de Exce-lencia "María de Maeztu"- Centro de Astrobiología (INTA-CSIC)D.D. acknowledges support from the TESS Guest Investigator Programgrant 80NSSC19K1727 and NASA Exoplanet Research Program grantMNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 exoplanet (Foreman-Mackey et al. 2020) and its dependencies (Agol et al. 2019;Astropy Collaboration et al. 2013, 2018; Foreman-Mackey et al. 2020;Foreman-Mackey et al. 2017; Foreman-Mackey 2018; Luger et al. 2019;Salvatier et al. 2016; Theano Development Team 2016); numpy (Harriset al. 2020); scipy (Virtanen et al. 2020); pandas (McKinney et al.2011); astropy (Robitaille et al. 2013); matplotlib (Hunter 2007); AstroImageJ (Collins et al. 2017),;
TAPIR (Jensen 2013).
The data underlying this article is publicly available -
TESS datais stored on the Mikulski Archive for Space Telescopes (MAST) at https://archive . stsci . edu/tess/ , while HARPS data is bothavailable on the Data & Analysis Center for Exoplanets (DACE) at https://dace . unige . ch/ , and in Appendix tables B2 & B3. REFERENCES
Adibekyan V. Z., Sousa S. G., Santos N. C., Delgado Mena E., GonzálezHernández J. I., Israelian G., Mayor M., Khachatryan G., 2012, A&A,545, A32Adibekyan V., et al., 2015, A&A, 583, A94Agol E., Luger R., Foreman-Mackey D., 2019, arXiv e-printsAigrain S., Pont F., Zucker S., 2012, Monthly Notices of the Royal Astro-nomical Society, 419, 3147Akeson R., et al., 2013, Publications of the Astronomical Society of thePacific, 125, 989Aller A., Lillo-Box J., Jones D., Miranda L. F., Barceló Forteza S., 2020,A&A, 635, A128Armstrong D. J., et al., 2020, Nature, 583, 39Astropy Collaboration et al., 2013, A&A, 558, A33Astropy Collaboration et al., 2018, AJ, 156, 123Baranne A., et al., 1996, A&AS, 119, 373Barros S. C. C., Demangeon O., Díaz R. F., Cabrera J., Santos N. C., FariaJ. P., Pereira F., 2020, A&A, 634, A75Bean J. L., Raymond S. N., Owen J. E., 2020, arXiv e-prints, p.arXiv:2010.11867Beaugé C., Nesvorný D., 2012, ApJ, 751, 119Beichman C., et al., 2014, PASP, 126, 1134Benz W., et al., 2020, arXiv preprint arXiv:2009.11633Berger T. A., Huber D., van Saders J. L., Gaidos E., Tayar J., Kraus A. L.,2020, AJ, 159, 280Bertran de Lis S., Delgado Mena E., Adibekyan V. Z., Santos N. C., SousaS. G., 2015b, A&A, 576, A89Bertran de Lis S., Delgado Mena E., Adibekyan V. Z., Santos N. C., SousaS. G., 2015a, A&A, 576, A89Bitsch B., Battistini C., 2020, Astronomy & Astrophysics, 633, A10Boisse I., et al., 2009, A&A, 495, 959Boisse I., Bouchy F., Hébrard G., Bonfils X., Santos N., Vauclair S., 2011,Astronomy & Astrophysics, 528, A4Bolmont E., Raymond S. N., Leconte J., Matt S. P., 2012, Astronomy &Astrophysics, 544, A124Brown T. M., et al., 2013, Publications of the Astronomical Society of thePacific, 125, 1031Brown A., et al., 2018, Astronomy & astrophysics, 616, A1 Buchner J., et al., 2014, A&A, 564, A125Buchschacher N., Ségransan D., Udry S., Díaz R., 2015, in Taylor A. R.,Rosolowsky E., eds, Astronomical Society of the Pacific Conference Se-ries Vol. 495, Astronomical Data Analysis Software an Systems XXIV(ADASS XXIV). p. 7Carleo I., et al., 2020, arXiv preprint arXiv:2004.10095Carrera D., Ford E. B., Izidoro A., 2019, MNRAS, 486, 3874Choi J., Dotter A., Conroy C., Cantiello M., Paxton B., Johnson B. D., 2016,ApJ, 823, 102Claret A., 2017, Astronomy & Astrophysics, 600, A30Cloutier R., et al., 2020, The Astronomical Journal, 160, 3Collins K. A., Kielkopf J. F., Stassun K. G., Hessman F. V., 2017, AJ, 153,77Costa Silva A. R., Delgado Mena E., Tsantaki M., 2020, A&A, 634, A136Cresswell P., Nelson R. P., 2006, A&A, 450, 833Deck K. M., Agol E., 2015, The Astrophysical Journal, 802, 116Delgado Mena E., Israelian G., González Hernández J. I., Bond J. C., SantosN. C., Udry S., Mayor M., 2010, ApJ, 725, 2349Delgado Mena E., et al., 2014, A&A, 562, A92Delgado Mena E., Tsantaki M., Adibekyan V. Z., Sousa S. G., Santos N. C.,González Hernández J. I., Israelian G., 2017, A&A, 606, A94Delgado Mena E., et al., 2019, A&A, 624, A78Dorn C., Khan A., Heng K., Connolly J. A., Alibert Y., Benz W., TackleyP., 2015, Astronomy & Astrophysics, 577, A83Dos Santos L. A., et al., 2016, Astronomy & Astrophysics, 592, A156Dumusque X., et al., 2019, A&A, 627, A43Espinoza N., 2018, arXiv preprint arXiv:1811.04859Figueira P., et al., 2012, A&A, 541, A139Ford E. B., Rasio F. A., 2008, ApJ, 686, 621Foreman-Mackey D., 2018, Research Notes of the American AstronomicalSociety, 2, 31Foreman-Mackey D., Agol E., Ambikasaran S., Angus R., 2017, AJ, 154,220Foreman-Mackey D., Luger R., Czekala I., Agol E., Price-Whelan A., Barclay T., 2020, exoplanet-dev/exoplanet v0.3.2,doi:10.5281/zenodo.1998447, https://doi . org/10 . . Franchini M., et al., 2020, ApJ, 888, 55Fulton B. J., et al., 2017, The Astronomical Journal, 154, 109Ginzburg S., Schlichting H. E., Sari R., 2018, Monthly Notices of the RoyalAstronomical Society, 476, 759Gomes da Silva J., Santos N. C., Bonfils X., Delfosse X., Forveille T., UdryS., 2011, A&A, 534, A30Greene T. P., Line M. R., Montero C., Fortney J. J., Lustig-Yaeger J., LutherK., 2016, ApJ, 817, 17Grunblatt S. K., Howard A. W., Haywood R. D., 2015, The AstrophysicalJournal, 808, 127Harris C. R., et al., 2020, Nature, 585, 357Haywood R. D., et al., 2014, MNRAS, 443, 2517Henden A. A., Levine S., Terrell D., Welch D. L., 2015, in AmericanAstronomical Society Meeting Abstracts 000
Adibekyan V. Z., Sousa S. G., Santos N. C., Delgado Mena E., GonzálezHernández J. I., Israelian G., Mayor M., Khachatryan G., 2012, A&A,545, A32Adibekyan V., et al., 2015, A&A, 583, A94Agol E., Luger R., Foreman-Mackey D., 2019, arXiv e-printsAigrain S., Pont F., Zucker S., 2012, Monthly Notices of the Royal Astro-nomical Society, 419, 3147Akeson R., et al., 2013, Publications of the Astronomical Society of thePacific, 125, 989Aller A., Lillo-Box J., Jones D., Miranda L. F., Barceló Forteza S., 2020,A&A, 635, A128Armstrong D. J., et al., 2020, Nature, 583, 39Astropy Collaboration et al., 2013, A&A, 558, A33Astropy Collaboration et al., 2018, AJ, 156, 123Baranne A., et al., 1996, A&AS, 119, 373Barros S. C. C., Demangeon O., Díaz R. F., Cabrera J., Santos N. C., FariaJ. P., Pereira F., 2020, A&A, 634, A75Bean J. L., Raymond S. N., Owen J. E., 2020, arXiv e-prints, p.arXiv:2010.11867Beaugé C., Nesvorný D., 2012, ApJ, 751, 119Beichman C., et al., 2014, PASP, 126, 1134Benz W., et al., 2020, arXiv preprint arXiv:2009.11633Berger T. A., Huber D., van Saders J. L., Gaidos E., Tayar J., Kraus A. L.,2020, AJ, 159, 280Bertran de Lis S., Delgado Mena E., Adibekyan V. Z., Santos N. C., SousaS. G., 2015b, A&A, 576, A89Bertran de Lis S., Delgado Mena E., Adibekyan V. Z., Santos N. C., SousaS. G., 2015a, A&A, 576, A89Bitsch B., Battistini C., 2020, Astronomy & Astrophysics, 633, A10Boisse I., et al., 2009, A&A, 495, 959Boisse I., Bouchy F., Hébrard G., Bonfils X., Santos N., Vauclair S., 2011,Astronomy & Astrophysics, 528, A4Bolmont E., Raymond S. N., Leconte J., Matt S. P., 2012, Astronomy &Astrophysics, 544, A124Brown T. M., et al., 2013, Publications of the Astronomical Society of thePacific, 125, 1031Brown A., et al., 2018, Astronomy & astrophysics, 616, A1 Buchner J., et al., 2014, A&A, 564, A125Buchschacher N., Ségransan D., Udry S., Díaz R., 2015, in Taylor A. R.,Rosolowsky E., eds, Astronomical Society of the Pacific Conference Se-ries Vol. 495, Astronomical Data Analysis Software an Systems XXIV(ADASS XXIV). p. 7Carleo I., et al., 2020, arXiv preprint arXiv:2004.10095Carrera D., Ford E. B., Izidoro A., 2019, MNRAS, 486, 3874Choi J., Dotter A., Conroy C., Cantiello M., Paxton B., Johnson B. D., 2016,ApJ, 823, 102Claret A., 2017, Astronomy & Astrophysics, 600, A30Cloutier R., et al., 2020, The Astronomical Journal, 160, 3Collins K. A., Kielkopf J. F., Stassun K. G., Hessman F. V., 2017, AJ, 153,77Costa Silva A. R., Delgado Mena E., Tsantaki M., 2020, A&A, 634, A136Cresswell P., Nelson R. P., 2006, A&A, 450, 833Deck K. M., Agol E., 2015, The Astrophysical Journal, 802, 116Delgado Mena E., Israelian G., González Hernández J. I., Bond J. C., SantosN. C., Udry S., Mayor M., 2010, ApJ, 725, 2349Delgado Mena E., et al., 2014, A&A, 562, A92Delgado Mena E., Tsantaki M., Adibekyan V. Z., Sousa S. G., Santos N. C.,González Hernández J. I., Israelian G., 2017, A&A, 606, A94Delgado Mena E., et al., 2019, A&A, 624, A78Dorn C., Khan A., Heng K., Connolly J. A., Alibert Y., Benz W., TackleyP., 2015, Astronomy & Astrophysics, 577, A83Dos Santos L. A., et al., 2016, Astronomy & Astrophysics, 592, A156Dumusque X., et al., 2019, A&A, 627, A43Espinoza N., 2018, arXiv preprint arXiv:1811.04859Figueira P., et al., 2012, A&A, 541, A139Ford E. B., Rasio F. A., 2008, ApJ, 686, 621Foreman-Mackey D., 2018, Research Notes of the American AstronomicalSociety, 2, 31Foreman-Mackey D., Agol E., Ambikasaran S., Angus R., 2017, AJ, 154,220Foreman-Mackey D., Luger R., Czekala I., Agol E., Price-Whelan A., Barclay T., 2020, exoplanet-dev/exoplanet v0.3.2,doi:10.5281/zenodo.1998447, https://doi . org/10 . . Franchini M., et al., 2020, ApJ, 888, 55Fulton B. J., et al., 2017, The Astronomical Journal, 154, 109Ginzburg S., Schlichting H. E., Sari R., 2018, Monthly Notices of the RoyalAstronomical Society, 476, 759Gomes da Silva J., Santos N. C., Bonfils X., Delfosse X., Forveille T., UdryS., 2011, A&A, 534, A30Greene T. P., Line M. R., Montero C., Fortney J. J., Lustig-Yaeger J., LutherK., 2016, ApJ, 817, 17Grunblatt S. K., Howard A. W., Haywood R. D., 2015, The AstrophysicalJournal, 808, 127Harris C. R., et al., 2020, Nature, 585, 357Haywood R. D., et al., 2014, MNRAS, 443, 2517Henden A. A., Levine S., Terrell D., Welch D. L., 2015, in AmericanAstronomical Society Meeting Abstracts 000 , 1–15 (2021) H.P. Osborn et al.
Kempton E. M.-R., et al., 2018, Publications of the Astronomical Societyof the Pacific, 130, 114401King G. W., Wheatley P. J., Bourrier V., Ehrenreich D., 2019, MonthlyNotices of the Royal Astronomical Society: Letters, 484, L49Kipping D. M., 2013a, Monthly Notices of the Royal Astronomical Society:Letters, 434, L51Kipping D. M., 2013b, Monthly Notices of the Royal Astronomical Society,435, 2152Kurucz R. L., 1993, SYNTHE spectrum synthesis programs and line dataLi J., Tenenbaum P., Twicken J. D., Burke C. J., Jenkins J. M., QuintanaE. V., Rowe J. F., Seader S. E., 2019, PASP, 131, 024506Lissauer J. J., et al., 2011, The Astrophysical Journal Supplement Series,197, 8Lodders K., Palme H., Gail H.-P., 2009, in , Solar system. Springer, pp712–770Lomb N. R., 1976, Ap&SS, 39, 447Luger R., Agol E., Foreman-Mackey D., Fleming D. P., Lustig-Yaeger J.,Deitrick R., 2019, AJ, 157, 64Martinez C. F., Cunha K., Ghezzi L., Smith V. V., 2019, The AstrophysicalJournal, 875, 29Maxted P. F. L., et al., 2011, PASP, 123, 547Mayor M., et al., 2003, The Messenger, 114, 20McCully C., Volgenau N. H., Harbeck D.-R., Lister T. A., Saunders E. S.,Turner M. L., Siiverd R. J., Bowman M., 2018, in Proc. SPIE. p.107070K ( arXiv:1811.04163 ), doi:10.1117/12.2314340McKinney W., et al., 2011, Python for High Performance and ScientificComputing, 14Milbourne T., et al., 2019, The Astrophysical Journal, 874, 107Mortier A., Collier Cameron A., 2017, A&A, 601, A110Mortier A., Faria J. P., Correia C. M., Santerne A., Santos N. C., 2015,A&A, 573, A101Mousis O., Deleuil M., Aguichine A., Marcq E., Naar J., Aguirre L. A.,Brugger B., Goncalves T., 2020, arXiv preprint arXiv:2002.05243Mulders G. D., Pascucci I., Apai D., Frasca A., Molenda-Żakowicz J., 2016,The Astronomical Journal, 152, 187Nielsen L. D., et al., 2020, Monthly Notices of the Royal AstronomicalSociety, 492, 5399Niraula P., et al., 2017, The Astronomical Journal, 154, 266Nowak G., et al., 2020, arXiv preprint arXiv:2003.01140Onken C. A., et al., 2020, arXiv e-prints, p. arXiv:2008.10359Otegi J., Bouchy F., Helled R., 2020, Astronomy & Astrophysics, 634, A43Owen J. E., Wu Y., 2017, The Astrophysical Journal, 847, 29Pepe F., et al., 2002, The Messenger, 110, 9Pollacco D. L., et al., 2006, PASP, 118, 1407Queloz D., et al., 2001, A&A, 379, 279Quinn S. N., et al., 2019, The Astronomical Journal, 158, 177Ricker G. R., et al., 2016, in Space Telescopes and Instrumentation 2016:Optical, Infrared, and Millimeter Wave. p. 99042BRobitaille T. P., et al., 2013, Astronomy & Astrophysics, 558, A33Robles J. A., Lineweaver C. H., Grether D., Flynn C., Egan C. A., PracyM. B., Holmberg J., Gardner E., 2008, ApJ, 684, 691Rogers J. G., Owen J. E., 2020, arXiv preprint arXiv:2007.11006Salvatier J., Wiecki T. V., Fonnesbeck C., 2016, PeerJ Computer Science, 2,e55Santos N. C., et al., 2013a, A&A, 556, A150Santos N. C., et al., 2013b, A&A, 556, A150Scargle J. D., 1982, ApJ, 263, 835Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525Skrutskie M., et al., 2006, The Astronomical Journal, 131, 1163Smith J. C., et al., 2012, PASP, 124, 1000Sneden C., 1973, ApJ, 184, 839Sousa S. G., et al., 2008, A&A, 487, 373Sousa S. G., Santos N. C., Israelian G., Mayor M., Udry S., 2011, A&A,533, A141Stassun K. G., Torres G., 2016, AJ, 152, 180Stassun K. G., Torres G., 2018, ApJ, 862, 61Stassun K. G., Collins K. A., Gaudi B. S., 2017, AJ, 153, 136Stassun K. G., Corsaro E., Pepper J. A., Gaudi B. S., 2018, AJ, 155, 22 Stassun K. G., et al., 2019, The Astronomical Journal, 158, 138Stumpe M. C., et al., 2012, Publications of the Astronomical Society of thePacific, 124, 985Stumpe M. C., Smith J. C., Catanzarite J. H., Van Cleve J. E., Jenkins J. M.,Twicken J. D., Girouard F. R., 2014, Publications of the AstronomicalSociety of the Pacific, 126, 100Suárez Mascareño A., et al., 2020, A&A, 639, A77Theano Development Team 2016, arXiv e-prints, abs/1605.02688Tokovinin A., 2018, Publications of the Astronomical Society of the Pacific,130, 035002Torres G., Andersen J., Giménez A., 2010, A&ARv, 18, 67Twicken J. D., et al., 2018, PASP, 130, 064502Van Eylen V., Agentoft C., Lundkvist M., Kjeldsen H., Owen J. E., Ful-ton B. J., Petigura E., Snellen I., 2018, Monthly Notices of the RoyalAstronomical Society, 479, 4786Vanderburg A., Plavchan P., Johnson J. A., Ciardi D. R., Swift J., Kane S. R.,2016, Monthly Notices of the Royal Astronomical Society, 459, 3565Virtanen P., et al., 2020, Nature methods, 17, 261Winn J. N., Fabrycky D. C., 2015, Annual Review of Astronomy and Astro-physics, 53Wright N. J., Newton E. R., Williams P. K., Drake J. J., Yadav R. K., 2018,Monthly Notices of the Royal Astronomical Society, 479, 2351Zeng L., et al., 2019, Proceedings of the National Academy of Sciences,116, 9723Ziegler C., Tokovinin A., Briceno C., Mang J., Law N., Mann A., 2020,yCat, pp J–AJ
APPENDIX A: AUTHOR AFFILIATIONS NCCR/PlanetS, Centre for Space & Habitability, University ofBern, Bern, Switzerland Department of Physics and Kavli Institute for Astrophysicsand Space Research, Massachusetts Institute of Technology,Cambridge, MA 02139, USA Centre for Exoplanets and Habitability, University of Warwick,Gibbet Hill Road, Coventry, CV4 7AL, UK Department of Physics, University of Warwick, Gibbet Hill Road,Coventry CV4 7AL, UK Instituto de Astrofísica e Ciências do Espaço, Universidade doPorto, CAUP, Rua das Estrelas, 4150-762 Porto, Portugal Harvard-Smithsonian Center for Astrophysics, 60 Garden St,Cambridge, MA, 02138, USA NASA Ames Research Center,Moffett Field, CA 94035, USA Astrophysics Group, Keele University, Staffs ST5 5BG, U.K. Centro de Astrobiologiía (CAB,CSIC-INTA), Dep. de Astrofísica,ESAC campus, 28692, Villanueva de la Cañada, Madrid, Spain Geneva Observatory, University of Geneva, Chemin desMailettes 51, 1290 Versoix, Switzerland Departamento de Física e Astronomia, Faculdade de Ciências,Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto,Portugal Dunlap Institute for Astronomy and Astrophysics, University ofToronto, 50 St. George Street, Toronto, Ontario M5S 3H4, Canada Cerro Tololo Inter-American Observatory, Casilla 603, LaSerena, Chile SETI Institute American Association of Variable Star Observers, 49 Bay StateRoad, Cambridge, MA 02138, USA International Center for Advanced Studies (ICAS) and ICIFI(CONICET), ECyT-UNSAM, Campus Miguelete, 25 de Mayo yFrancia, (1650) Buenos Aires, Argentina. Department of Physics and Astronomy, University of NewMexico, 1919 Lomas Blvd NE, Albuquerque, NM 87131, USA
MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 Aix Marseille Univ, CNRS, CNES, LAM, Marseille, France Institute for Computational Science, University ofZurich,Winterthurerstr. 190, CH-8057 Zurich, Switzerland Kavli Fellow European Southern Observatory, Alonso de Cordova 3107,Vitacura, Santiago, Chile Department of Physics & Astronomy, Swarthmore College,Swarthmore PA 19081, USA Department of Physics and Astronomy, The University of NorthCarolina at Chapel Hill, Chapel Hill, NC 27599-3255, USA Department of Astronomy, University of Maryland, CollegePark, MD 20742, USA Space Telescope Science Institute, 3700 San Martin Drive,Baltimore, MD 21218, USA Department of Earth, Atmospheric and Planetary Sciences,Massachusetts Institute of Technology, Cambridge, MA 02139,USA Department of Physics & Astronomy, Vanderbilt University,6301 Stevenson Center Ln., Nashville, TN 37235, USA Department of Astrophysical Sciences, Princeton University, 4Ivy Lane, Princeton, NJ 08544, USA
APPENDIX B: EXTRA TABLES
This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000
This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000 , 1–15 (2021) H.P. Osborn et al.
Table B1.
List of free parameters used in the exoplanet combined analysis of the
TESS light curve and HARPS radial velocities with their associated priorand posterior distributions.Parameter Prior Posterior
Stellar parameters
Stellar surface temperature, 𝑇 eff [K] N ( . , . ) ± 𝑀 𝑠 [ 𝑀 (cid:12) ] N ( . , . ) . ± . 𝑅 𝑠 [ 𝑅 (cid:12) ] N ( . , . ) . ± . Orbital parameters
Transit Epoch, 𝑡 [BJD-2457000] b N ( . , . ) . + . − . Transit Epoch, 𝑡 [BJD-2457000] c N ( . , . ) . ± . 𝑃 [d] b N U ( . , . , . , . ) . + . − . Orbital Period, 𝑃 [d] c N U ( . , . , . , . ) . + . − . Orbital Eccentricity, 𝑒 b 𝛽 ( . . ) 𝑎 . + . − . Orbital Eccentricity, 𝑒 c 𝛽 ( . . ) 𝑎 . + . − . Argument of periastron, Ω b U (− 𝜋, 𝜋 ) 𝑏 − . ± . Ω c U (− 𝜋, 𝜋 ) 𝑏 − . ± . Photometric parameters log radius ratio [log 𝑅 𝑝 / 𝑅 𝑠 ] b U (− . , − . ) − . ± . U ( , + 𝑅 𝑝 / 𝑅 𝑠 ) 𝑐 . ± . 𝑎 LD N U ( . , . , . , . ) . ± . 𝑏 LD N U ( . , . , . , . ) . + . − . Photometric jitter [log ppt]
N ( . , . ) − . + . − . Photometric GP power
I ( . , . ) 𝑑 . ± . 𝑑 − ] I ( . , . ) 𝑑 . + . − . Photometric GP mean [ppt]
I ( . , . ) 𝑑 . + . − . HARPS parameterslog RV semi-amplitude, log 𝐾 b N ( . , . ) . ± . 𝐾 c N ( . , . ) . ± . − ] N ( . , . ) . ± . − 𝑑 − ] N ( . , . ) − . ± . − ] N ( . , . ) − . + . − . HARPS log jitter S index
N ( . 𝑒 − , . ) − . + . − . HARPS log jitter FWHM [m s − ] N ( . , . ) . ± . N ( . , . ) − . + . − . HARPS mean FWHM [m s − ] N ( . , . ) . ± . N ( . , . ) . + . − . HARPS GP log amplitude S-index
N (− . , . ) − . + . − . HARPS GP log amplitude FWHM
N ( . , . ) . ± . 𝑃 rot / log 𝑑 N U ( . , . , . , . ) . ± . 𝑄 N ( . , . ) − . + . − . HARPS GP log quality differential, Δ 𝑄 N ( . , . ) . + . − . HARPS GP 𝑃 rot - 𝑃 rot / U ( , ) . ± . N ( 𝜇 ; 𝜎 ) is a normal distribution with mean 𝜇 and width 𝜎 , U ( 𝑎 ; 𝑏 ) is a uniform distribution between 𝑎 and 𝑏 , N U ( 𝜇 ; 𝜎 , 𝑎, 𝑏 ) is a normal distribution with mean 𝜇 and width 𝜎 multiplied with a uniform distribution between 𝑎 and 𝑏 , 𝛽 ( 𝑎 ; 𝑏 ) is a Beta distribution with parameters 𝑎 and 𝑏 , and I ( 𝜇 ; 𝜎 ) is a distribution directly interpolated fromthe output of a pre-trained distribution with mean 𝜇 and standard deviation 𝜎 (although the distribution may not follow anormal distribution). Posterior values and uncertainties represent the median and 1 𝜎 error boundaries. All other values (e.g.presented in Table 3) are directly determined from these fitted quantities. The prior uncertainties of input parameters 𝑡 and 𝑃 were inflated from the input data uncertainties by factors of: 𝑡 ,𝑏 = × , 𝑡 ,𝑐 = × , 𝑃 𝑏 = × , 𝑃 𝑐 = × . 𝑎 Describedin Kipping (2013a). 𝑏 Reparameterised in exoplanet to avoid discontinuities at ± 𝜋 . 𝑐 exoplanet reparameterization ofEspinoza (2018). 𝑑 PyMc3
Interpolation function of pre-trained GP. MNRAS , 1–15 (2021) hot mini-Neptune orbiting HD 110113 Table B2.
HARPS spectroscopy from first season (June - August 2019)
Time RV 𝜎 RV 𝑆 MW 𝜎 𝑆 FWHM 𝜎 FWHM [ BJD − ] [ms − ] – [ms − ] . . .
95 0 . .
004 7281 . . . .
32 2 . . . . . . − .
08 2 .
39 0 . . . . . − .
07 1 .
44 0 . . . . . − .
37 1 .
49 0 .
011 0 . . . . − .
78 1 . − . . . . . .
04 1 . − . .
003 7283 . . . − . . − . . . . . − .
17 1 . − . .
004 7276 . . . − .
22 1 . − . . . . . − .
01 1 . − . . . . . .
21 1 . − . . . . . . . − . . . . . − .
47 1 . − . . . . . − .
81 1 . − .
015 0 . . . . − .
35 1 . − . . . . . − .
13 1 . − . . . . . .
86 1 .
53 0 . . . . . − .
04 1 . − . . . . . . .
62 0 . . . . . − .
22 1 . − . . . . . − .
59 1 .
43 0 . . . . . − . . − . . . . . − .
66 1 . − .
001 0 . . . . − .
58 1 . − . . . . . − .
82 1 . − . . . . . − .
88 1 . − . . . . . − .
52 1 . − . . . . . − .
48 1 . − . . . . . − .
11 1 . − . .
002 7278 . . . − .
81 1 . − .
016 0 . . . . − .
71 1 . − . . . . . − .
15 1 . − . . . . . − .
64 1 . − . .
003 7279 . . . − .
43 1 . − . . . . . .
68 1 . − . . . . . .
35 1 . − .
019 0 . . . . .
11 1 . − . . . . . .
79 1 . − .
005 0 . . . . .
87 1 . − . . . . . .
23 1 . − . . . . . .
52 1 .
36 0 .
002 0 . . . . .
59 1 .
19 0 . .
002 7288 . . .
462 6 .
68 1 .
21 0 . . . . . .
74 2 . − . . . . . − .
52 1 .
57 0 . . . . . − .
82 1 . − . . . . . − .
25 1 .
29 0 . . . . . − .
83 1 . − . . . . . − .
04 1 . − . . . . . − .
62 1 . − . . . . . − .
69 2 . − .
024 0 . . . . − .
55 1 . − . . . . . − .
22 1 . − . . . . . − .
27 2 . − . . . . . .
39 1 . − . . . . Table B3.
HARPS spectroscopy from second season (Dec 2019 - Feb 2020).
Time RV 𝜎 RV 𝑆 MW 𝜎 𝑆 FWHM 𝜎 FWHM [ BJD − ] [ms − ] – [ms − ] . .
17 1 . . . . . . .
52 1 .
26 0 . . . . . .
27 1 .
15 0 . . . . . .
16 1 .
19 0 . . . . . − .
96 1 .
45 0 . . . . . − .
03 1 . − . . . . . − .
35 1 . − . . . . . − .
38 1 . − . .
002 7272 . . . − .
59 1 . − . . . . . − .
97 1 . − . . . . . − .
28 1 . − . . . . . .
75 1 .
08 0 . . . . . .
51 1 . − .
001 0 . . . . − .
78 1 .
17 0 . . . . . − .
05 1 .
42 0 . . . . . .
87 1 .
25 0 . . . . . .
84 1 .
68 0 . .
003 7298 . . .
757 4 .
48 1 . − . . . . . − .
07 1 . − . . . . . .
95 1 .
38 0 . . . . .
775 7 .
83 1 .
33 0 . .
002 7300 . . . .
32 1 .
31 0 . . . . . .
81 1 .
44 0 . . . . . − . .
26 0 .
008 0 . . . . − .
61 1 .
19 0 . . . . . .
82 1 .
22 0 . . . . .
85 3 . .
22 0 . . . . . − .
94 1 . . . . . . − .
75 1 .
19 0 . . . . . .
21 1 .
13 0 . . . . . . .
06 0 . . . . . − .
21 1 .
25 0 . . . . . − .
66 1 .
46 0 .
005 0 . . . . − .
78 1 .
41 0 . .
002 7289 . . . .
07 1 .
26 0 . . . . . .
09 1 . − . . . . . . .
25 0 . . . . . .
46 1 .
56 0 . . . . . .
05 1 .
22 0 . . . . . .
67 1 .
22 0 . . . . . .
47 1 .
23 0 . . . . . .
18 1 .
23 0 . . . . . .
27 1 .
17 0 . . . . . .
62 1 .
26 0 . . . . . − .
86 1 . − . . . . . − .
95 1 . − . . . . . . . − . . . . . .
79 1 . − . . . . . . .
18 0 . . . . . .
63 1 .
17 0 . . . . . .
06 1 . . . . . . .
46 1 .
11 0 . . . . . .
59 1 .
21 0 . .
002 7289 . . . .
17 1 .
07 0 . . . . . . .
13 0 . . . . . .
72 1 .
16 0 . . . . . .
72 1 .
07 0 .
011 0 . . . .
885 15 . .
26 0 . . . . MNRAS000