The GAPS Programme with HARPS-N at TNG Based on observations collected at the Italian Telescopio Nazionale Galileo (TNG), operated on the island of La Palma by the Fundación Galileo Galilei of the Istituto Nazionale di Astrofisica (INAF) at the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofísica de Canarias, in the frame of the programme Global Architecture of Planetary Systems (GAPS).

The GAPS Programme with HARPS-N at TNG. III: The retrograde orbit of HAT-P-18b

III: The retrograde orbit of HAT-P-18b
Key Words.:
Planetary systems – Techniques: spectroscopic, radial velocities – Stars: individual: HAT-P-18

The measurement of the Rossiter-McLaughlin effect for transiting exoplanets places constraints on the orientation of the orbital axis with respect to the stellar spin axis, which can shed light on the mechanisms shaping the orbital configuration of planetary systems. Here we present the interesting case of the Saturn-mass planet HAT-P-18b, which orbits one of the coolest stars for which the Rossiter-McLaughlin effect has been measured so far. We acquired a spectroscopic time-series, spanning a full transit, with the HARPS-N spectrograph mounted at the TNG telescope. The very precise radial velocity measurements delivered by the HARPS-N pipeline were used to measure the Rossiter-McLaughlin effect. Complementary new photometric observations of another full transit were also analysed to obtain an independent determination of the star and planet parameters. We find that HAT-P-18b lies on a counter-rotating orbit, the sky-projected angle between the stellar spin axis and the planet orbital axis being deg. By joint modelling of the radial velocity and photometric data we obtain new determinations of the star ( M; R; km s) and planet ( M; R) parameters. Our spectra provide for the host star an effective temperature K, a surface gravity of cm s, and an iron abundance of [Fe/H] = . HAT-P-18b is one of the few planets known to transit a star with K on a retrograde orbit. Objects such as HAT-P-18b (low planet mass and/or relatively long orbital period) most likely have a weak tidal coupling with their parent stars, therefore their orbits preserve any original misalignment. As such, they are ideal targets to study the causes of orbital evolution in cool main-sequence stars.

1 Introduction

The number of known extrasolar planets has recently passed the milestone of one thousand. While many discovery surveys are still ongoing, the characterization of known extrasolar planetary systems is gaining ever more attention. Transiting extrasolar planets (TEPs) are especially interesting as they allow for the direct determination of fundamental parameters such as planetary mass and radius (Southworth, 2012). Moreover, observations of secondary eclipses put constraints on the planet albedo and brightness temperature, while transmission spectroscopy can be used to probe molecular and atomic features in the planet atmospheres.

Another possibility offered by TEPs is to study the Rossiter-McLaughlin (RM) effect, which is an anomaly in the radial velocity orbital trend that occurs when the planet moves across the stellar photospheric disc (see Hirano et al. 2011 and references therein). The measurement of the RM effect permits the determination of the angle , the projection on the sky plane of the misalignment angle between the stellar spin axis and the planet orbital axis. The knowledge of can give insight into the mechanisms of formation and orbital migration of exoplanets (Naoz et al. 2011; Nagasawa et al. 2008; Wu & Lithwick 2011).

In the context of GAPS, a long-term observational programme with HARPS-N at TNG (Covino et al. 2013, hereafter Paper I; Desidera et al. 2013), we are carrying out a sub-programme aimed at measuring the RM effect in a sample of TEP host stars. We plan to explore a wide assortment of stellar temperatures, ages, and masses, as well as diverse orbital (period, eccentricity) and physical (mass, radius) planet properties.

In this paper, we report on the measurement of the RM effect for the HAT-P-18 transiting system (Hartman et al., 2011). HAT-P-18b is a Saturn-mass planet orbiting a K2 dwarf star with a period days. Hartman et al. (2011) (hereafter H11) pointed out that with a density g cm, HAT-P-18b is not expected to have a significant heavy element core, according to the planetary models by Fortney et al. (2007).

2 Observations and data reduction

A time series of 20 spectra of HAT-P-18 was acquired with HARPS-N at TNG (Cosentino et al., 2012) in 2013 June 11-12, bracketing a full transit of HAT-P-18b. The exposure time was of 15 minutes, resulting in spectra with an S/N20 (per extracted pixel at 5500 Å), degrading to 15 for the last three spectra due to worsening seeing conditions (see Table 1). In the following months, three additional spectra were taken, two of which at phases nearly corresponding to the expected minimum and maximum of the radial velocity curve.

HARPS-N spectra were reduced using the standard pipeline. Radial velocities (RVs) were derived using the weighted cross-correlation function (CCF) method (Baranne et al. 1996, Pepe et al. 2002); for HAT-P-18 we used the K5 mask. The pipeline also provided rebinned 1D spectra that were used for the stellar atmosphere characterization (see Sec. 3.1).

A complete transit event of HAT-P-18b was observed on 2011 May 24, with the Bologna Faint Object Spectrograph & Camera (BFOSC) imager mounted on the 1.52-m Cassini Telescope at the Astronomical Observatory of Bologna in Loiano, Italy (Mancini et al., 2013; Ciceri et al., 2013). The night was not photometric and some data before ingress were rejected as they were affected by clouds. The CCD was used unbinned, giving a plate scale of for a total field-of-view of , and the transit was observed through a Gunn filter. The telescope was autoguided and defocussed to increase the exposure time to 140 sec, which minimisies the effects of systematic noises. The photometric data were derived using an upgraded version of the defot package (Southworth et al., 2009a, b).

BJD (TDB) RV error S/N2 3
[m s] [m s]
2456455.451721 11094.3 5.2 20.7 o
2456455.462430 11099.6 4.8 21.9 o
2456455.473147 11094.8 6.3 17.8 o
2456455.483869 11093.4 7.0 16.6 o
2456455.494596 11095.1 5.9 18.7 o
2456455.505314 11106.0 4.7 21.7 i
2456455.516031 11115.9 4.5 22.5 i
2456455.526749 11116.5 4.9 20.8 i
2456455.537462 11103.8 4.5 22.4 i
2456455.548184 11118.4 5.2 20.2 i
2456455.558915 11091.6 7.2 16.1 i
2456455.569633 11100.7 5.2 20.3 i
2456455.580346 11095.8 4.4 23.0 i
2456455.591059 11096.2 4.7 22.2 i
2456455.601781 11088.2 4.7 22.0 i
2456455.612495 11101.9 4.6 22.6 i
2456455.623212 11102.1 5.1 20.8 o
2456455.633939 11095.6 7.4 15.8 o
2456455.644665 11122.4 8.3 15.1 o
2456455.656845 11108.6 7.1 16.2 o
2456506.466435 11130.3 6.8 17.3 o
2456536.490373 11073.4 5.8 19.5 o
2456543.375155 11095.1 4.5 23.7 o
Table 1: HARPS-N RV measurements of HAT-P-18.

3 Results

3.1 Spectroscopic determination of stellar parameters

We derived the photospheric parameters of the planet-hosting star HAT-P-18 by applying two different methods on the mean of all the available HARPS-N spectra.

The first method relies on the use of the spectral analysis package MOOG (Sneden 1973, version 2013). As in Paper I, we measured the equivalent widths (EWs) of iron lines chosen from the list by Biazzo et al. (2012) and adopted the abfind driver within MOOG. We hence determined the effective temperature () by imposing that the Fe i abundance does not depend on the excitation potential of the lines, the microturbulence velocity () by imposing that the Fe i abundance is independent on the EW of the lines, and the surface gravity () by the Fe i/Fe ii ionization equilibrium condition. The projected rotational velocity  sin  was measured following the procedure described in D’Orazi et al. (2011).

The second method compares the composite HARPS-N spectrum with a grid of theoretical model spectra using spectral features that are sensitive to different photospheric parameters (Castelli & Kurucz 2004; Coelho et al. 2005; Gustafsson et al. 2008). Briefly, we used the wings of the Balmer lines to estimate the of the star, and the Mg i 5167, 5173, and 5184 Å, the Ca i 6162 and 6439 Å, and the Na i D lines to determine its . The iron abundance [Fe/H] and were derived by applying the method described in Blackwell & Shallis (1979) on isolated Fe i and Fe ii lines. The  sin  and macroturbulence velocity () were measured by fitting the profiles of several clean and isolated metal lines.

The two methods provided consistent results, well within the error bars. The final adopted values, obtained as the weighted mean of the two independent determinations, agree very well with the values by H11 (see Table 2). We note that the  sin  and are consistent within the errors with the values obtained by modelling the RM- and light-curve (see Sec. 3.2), thereby validating our global analysis.

3.2 RV and photometric data analysis

The RV and photometric data sets were analysed jointly. To this purpose we developed a MATLAB code that implements a global model and a data-fitting algorithm.

The model considers the parameters necessary to describe the planet and star position and velocity vectors at any given time, that is, the masses of the star and of the planet , the orbital period and eccentricity , the epoch and argument of periastron, the systemic RV ; the orbital space orientation is described by the inclination angle and the misalignment angle 5. Other parameters necessary to model the RM effect and the light curve are the stellar and planetary radius, the stellar projected rotational velocity , and the limb-darkening coefficients. Our model can implement each of the five equations proposed by Claret & Bloemen (2011) to describe the limb-darkening law. Other effects that can affect the measurements, such as stellar surface inhomogeneities, stellar differential rotation and convective blue-shift, are not included in the model. We refer to Paper I for the details of the method used to determine the RV anomaly when the planet is transiting the stellar disc.

The best-fit values of the parameters are obtained by a least-squares minimization algorithm. The region of the parameters space to be explored can be limited providing upper and lower limits to the parameter values. Most importantly, any number of linear and non-linear constraints can be set: this allows placing limits on other parameters (such as , , , see Table 2 for their definition), even though they are not direct parameters of the fit. The mass of the star, (), is preliminarily determined from evolutionary track models (Yi et al., 2001), adopting the values of the atmospheric parameters determined previously and using the value derived from the light curve. Evolutionary tracks also provide an estimate of the stellar age of 7.0 3.6 Gyr.

Together with our data sets, the global fit also considers the RVs presented in H11. We show in the top panel of Fig. 1 the phase-folded RV data, with superimposed the best-fit RV curve. We find the eccentricity to be , consistent with a circular orbit (Zakamska et al., 2011). The middle panel displays an expanded view of the phases around the transit. During the transit the RVs are first blue- and then red-shifted with respect to the orbital trend, indicating that the planet is moving on a retrograde orbit. The best-fit value for the sky-projected spin-orbit misalignment angle is deg. To evaluate the significance of the detection of the RM effect, we used the transit RV data alone (-0.050phase-0.024) and compared the values obtained by modelling the effect, =1.08, and by just fitting the orbital trend, =4.43. The bottom panel of Fig. 1 shows the phase-folded -band photometric data set and the best-fit light curve. We adopted the simple linear law to describe the stellar limb-darkening, as no significant improvement on the light curves fit is obtained by using a quadratic law.

The best-fit values for all the parameters are listed in Table 2, together with the errors that were determined by means of a Monte Carlo method. Our results agree well with those reported in H11.

Figure 1: Upper panel: phase-folded RV data set. Blue open circles are the HARPS-N data, black asterisks are the RVs from Hartman et al. (2011). Superimposed is the best-fit RV curve. Middle panel: zoom in the RV measurements spanning the transit. Lower panel: -band photometric time-series and best-fit light curve.
Parameter [Units] Value
Stellar spectra characterization
Effective temperature, [K]
Surface gravity, [cm s]
Iron abundance, [Fe/H]
Microturbulence, [km s]
Macroturbulence, [km s]
Proj. rot. vel., [km s]
RV and photometric data fit
Star mass, [M]
Planet mass, [M]
Star radius, [R]
Planet radius, [R]
Orbital period, [days]
Longitude of periastron, [deg]
Orbital inclination, [deg]
Epoch of periastron, [BJD]
Barycentric RV, [m s]
H11 RVs offset, [m s]
Proj. spin-orbit angle, [deg]
Proj. rot. vel., [km s]
Limb dark. — HARPS-N band,
Limb dark. —  band,
Normalized chi-square, 1.39
Derived parameters
Orbital semi-major axis, [AU]
Transit duration, [hours]
Impact parameter,
Transit depth —  band,
RV-curve semi-amplitude, [m s]
Star density, [g cm]
Star surface gravity, [cm s]
Planet density, [g cm]
Planet surface gravity, [cm s]
Planet equilibrium temperature, [K]
Table 2: Star and planet parameters of the system HAT-P-18

4 Discussion

Winn et al. (2010a) first noticed, and later Albrecht et al. (2012) (hereafter A12) confirmed, an empirical correlation between the spin-orbit relative orientation and the effective temperature of the host star: planets hosted by stars with 6250 K display a wide distribution of values, while planets around cooler stars are almost always well aligned (see Fig. 2).

This fact has been interpreted as supporting evidence of a scenario in which giant planets approach closely their parent stars following planet-planet gravitational scattering, Kozai-Lidov cycles, or secular chaotic orbital evolution, as opposed to gentle migration in a protoplanetary disc (Dawson & Murray-Clay 2013; Wu & Lithwick 2011; Baruteau et al. 2013). Initially, planets can have large misalignments around both cool and hot stars. Later on, because of their convective envelopes, tidal interactions are effective in cool stars to realign the systems on relatively short time-scales.

Figure 2: Compilation of the values of , measured via the RM effect, as a function of the host star effective temperature (see: HAT-P-18b is shown as a filled blue dot. For the two objects marked with red squares, the determination of is doubtful (see text for details). The red dashed line at = 6250 K marks the limit above which the mass of the convective shell becomes negligible (Winn et al., 2010a).

HAT-P-18, with a = K and a very large misalignment angle deg, seems to represent an exception in this scheme. However, as already argued by A12, the realignment time-scale also depends on other parameters such as the planet/star mass ratio and the scaled semi-major axis . In fact, using the equation (2) in A12 to estimate the alignment timescale , for HAT-P-18 we find yr, adopting the equilibrium tide theory of Zahn (1977). Given the present uncertainty of the tidal theory, rather than the absolute value of , it is interesting how it compares with the values for other TEPs. HAT-P-18b has one of the highest values of among the TEPs that orbit cool stars, inferior only to HAT-P-11b and HD80606b, which are also both significantly misaligned (see Fig. 24 in A12). Therefore, HAT-P-18b confirms that planets around cool stars can also have misaligned orbits, provided that they have a longer tidal realignment time-scale due to smaller mass (HAT-P-11b, Winn et al. 2010b; Kepler-63b, Sanchis-Ojeda et al. 2013) and/or larger orbital semi-major axis (WASP-8b, Queloz et al. 2010; HD80606b, Hébrard et al. 2010; Kepler-63b). The high misalignment angle deg of WASP-2b (Triaud et al., 2010) is at odds with its mass and separation (=0.9 M, =0.03 AU), but the reliability of the measure was confuted by Albrecht et al. (2011), based on new observations. The value of for WASP-80b (Triaud et al., 2013) is strongly dependent on the value assumed for , because of the nearly zero impact parameter.

The circularization of the orbit proceeds on a much shorter time-scale because of tidal dissipation inside the planet. Adopting a modified tidal quality factor of , corresponding to the value measured in Jupiter (Lainey et al., 2009), we obtain a damping time-scale of about 90 Myr for the eccentricity.

Rogers & Lin (2013) challenged the interpretation of A12 of the correlation. They advocated migration in the protoplanetary disc that produces aligned hot-Jupiters, and invoked a mechanism based on stellar internal gravity waves to explain the high obliquities found in hot stars. We point out that this mechanism is not applicable to stars such as HAT-P-18, which have a convective envelope.

5 Conclusions

We have found that the Saturn-mass planet hosted by HAT-P-18, a K2 dwarf star with K, lies on a retrograde orbit. We discussed how the existence of such object fits in the context of the current alternative theories of giant planet orbital migration. HAT-P-18b scores a point in favour of gravitational N-body (N3) interactions, while migration in the proto-planetary disc seems unable to explain its existence. HAT-P-18b, which is one of the very few planets around cool stars found to be on a retrograde orbit, also allows setting constraints on the efficiency of tidal interactions in obliquity damping.

ME acknowledges financial support from the Spanish Ministry project MINECO AYA2011-26244. We thank the TNG staff for help in the observations and with data retrieval from the TNG archive. The GAPS project in Italy acknowledges the support by INAF through the ”Progetti Premiali” funding scheme of the Italian Ministry of Education, University, and Research. DG acknowledges funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement n. 267251.


  1. thanks: Based on observations collected at the Italian Telescopio Nazionale Galileo (TNG), operated on the island of La Palma by the Fundación Galileo Galilei of the Istituto Nazionale di Astrofisica (INAF) at the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofísica de Canarias, in the frame of the programme Global Architecture of Planetary Systems (Gaps).
  2. per pixel at 5500 Å
  3. i  in-transit, o  out-of-transit
  4. The third angle, the longitude of the ascending node, is not considered as it does not affect the RV and photometric measurements; in other words, it is not an observable.


  1. Albrecht, S., Winn, J. N., Johnson, J. A., et al. 2011, ApJ, 738, 50
  2. Albrecht, S., Winn, J. N., Johnson, J. A., et al. 2012, ApJ, 757, 18
  3. Baranne, A., Queloz, D., Mayor, M., et al. 1996, A&AS, 119, 373
  4. Baruteau, C., Crida, A., Paardekooper, S.-J., et al. 2013, ArXiv e-prints
  5. Biazzo, K., D’Orazi, V., Desidera, S., et al. 2012, MNRAS, 427, 2905
  6. Blackwell, D. E. & Shallis, M. J. 1979, MNRAS, 186, 673
  7. Castelli, F. & Kurucz, R. L. 2004, ArXiv Astrophysics e-prints
  8. Ciceri, S., Mancini, L., Southworth, J., et al. 2013, A&A, 557, A30
  9. Claret, A. & Bloemen, S. 2011, A&A, 529, A75
  10. Coelho, P., Barbuy, B., Meléndez, J., Schiavon, R. P., & Castilho, B. V. 2005, A&A, 443, 735
  11. Cosentino, R., Lovis, C., Pepe, F., et al. 2012, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 8446
  12. Covino, E., Esposito, M., Barbieri, M., et al. 2013, A&A, 554, A28
  13. Dawson, R. I. & Murray-Clay, R. A. 2013, ApJ, 767, L24
  14. Desidera, S., Sozzetti, A., Bonomo, A. S., et al. 2013, A&A, 554, A29
  15. D’Orazi, V., Biazzo, K., & Randich, S. 2011, A&A, 526, A103
  16. Fortney, J. J., Marley, M. S., & Barnes, J. W. 2007, ApJ, 659, 1661
  17. Gustafsson, B., Edvardsson, B., Eriksson, K., et al. 2008, A&A, 486, 951
  18. Hartman, J. D., Bakos, G. Á., Sato, B., et al. 2011, ApJ, 726, 52
  19. Hébrard, G., Désert, J.-M., Díaz, R. F., et al. 2010, A&A, 516, A95
  20. Hirano, T., Suto, Y., Winn, J. N., et al. 2011, ApJ, 742, 69
  21. Lainey, V., Arlot, J.-E., Karatekin, Ö., & van Hoolst, T. 2009, Nature, 459, 957
  22. Mancini, L., Southworth, J., Ciceri, S., et al. 2013, A&A, 551, A11
  23. Nagasawa, M., Ida, S., & Bessho, T. 2008, ApJ, 678, 498
  24. Naoz, S., Farr, W. M., Lithwick, Y., Rasio, F. A., & Teyssandier, J. 2011, Nature, 473, 187
  25. Pepe, F., Mayor, M., Galland, F., et al. 2002, A&A, 388, 632
  26. Queloz, D., Anderson, D. R., Collier Cameron, A., et al. 2010, A&A, 517, L1
  27. Rogers, T. M. & Lin, D. N. C. 2013, ApJ, 769, L10
  28. Sanchis-Ojeda, R., Winn, J. N., Marcy, G. W., et al. 2013, ApJ, 775, 54
  29. Sneden, C. 1973, ApJ, 184, 839
  30. Southworth, J. 2012, MNRAS, 426, 1291
  31. Southworth, J., Hinse, T. C., Burgdorf, M. J., et al. 2009a, MNRAS, 399, 287
  32. Southworth, J., Hinse, T. C., Jørgensen, U. G., et al. 2009b, MNRAS, 396, 1023
  33. Triaud, A. H. M. J., Anderson, D. R., Collier Cameron, A., et al. 2013, A&A, 551, A80
  34. Triaud, A. H. M. J., Collier Cameron, A., Queloz, D., et al. 2010, A&A, 524, A25
  35. Winn, J. N., Fabrycky, D., Albrecht, S., & Johnson, J. A. 2010a, ApJ, 718, L145
  36. Winn, J. N., Johnson, J. A., Howard, A. W., et al. 2010b, ApJ, 723, L223
  37. Wu, Y. & Lithwick, Y. 2011, ApJ, 735, 109
  38. Yi, S., Demarque, P., Kim, Y.-C., et al. 2001, ApJS, 136, 417
  39. Zahn, J.-P. 1977, A&A, 57, 383
  40. Zakamska, N. L., Pan, M., & Ford, E. B. 2011, MNRAS, 410, 1895
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minumum 40 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description