Zodiacal Exoplanets in Time (ZEIT). XIV. He i Transit Spectroscopy of the 650 Myr Hyades Planet K2-136c

The radius distribution of sub-Jovian exoplanets is characterized by distinct populations of "sub-Neptunes" and "super-Earths," thought to reflect the presence or absence of a H/He-dominated envelope, separated by a sparsely occupied gap at 1.7–1.8R_⊕ (Fulton et al. 2017). Sub-Neptunes can transform into super-Earths by the loss of such envelopes and different mechanism for this loss (e.g., Owen 2019; Gupta & Schlichting 2021) can be tested by detection of an extended, escaping atmospheres among planets around young (≲1 Gyr) stars. The 121.4 nm Lyα resonant scattering line of H i is a sensitive of H-dominated atmospheres, but must be observed from space and suffers from attenuation by the interstellar matter and interference by geocoronal emission. In contrast, the 1083.2 nm line of metastable neutral orthohelium ("triplet" He i) offers a practical alternative since this line is readily accessible from the ground, but its production in a planetary atmosphere relies on irradiance by the host star (Oklopčić 2019).

K2-136 (LP 358-348) is a K5.5-type dwarf member of the 650 Myr old Hyades cluster. While ground-based monitoring failed to confirm a candidate transiting planet (Gaidos et al. 2014), monitoring by K2 during Campaign 13 revealed a system of three planets (Ciardi et al. 2018; Livingston et al. 2018; Mann et al. 2018): the middle and largest planet "c" has a radius of 2.9 ± 0.1R_⊕ and an orbital period of 17.3 days. This planet's radius and the suitability of K-type dwarfs for He i spectroscopy (Oklopčić 2019) motivated our observations. We obtained time-series spectra of the star during a 3.6 hr long transit on UT 2020 November 26. We used the IRD infrared echelle spectrograph on the 8.2 m Subaru telescope (λ/Δλ = 70,000, Kotani et al. 2018). The weather was partially cloudy, with occasional attenuation sufficient to cause the AO loop to open. The predicted transit center T_c occurred at 11:25 p.m. local time ±30 minutes (Mann et al. 2018). Observations began 3.15 hr before ingress and ended 1.16 hr after egress, with 19, 19, and 7 spectra obtained before, during, and after transit at an airmass between 1.03 and 1.78. Integration times were 420–600 s and the SNR was 30–90 per spectral pixel. A spectrum of the A0 star (HIP 26616) was obtained for telluric correction. Spectral images were processed and spectra extracted, blaze-corrected, and wavelength-calibrated using a custom reduction pipeline (Hirano et al. 2020b). Spectra were adjusted for barycenter motion (37.3 km s⁻¹ at T_c ). For spectra obtained during transit, a further Doppler shift to the planet rest frame was applied to any difference from the out-of-transit mean spectrum.

Figure 1 compares summed spectra obtained inside and outside of transit. This wavelength range includes stellar lines due to Si i and Na i as well as He i. The remaining features are variable telluric OH emission and H₂O absorption. These telluric interfere with the two redder components of the He i triplet, but the weaker single component is relatively free of contamination. This line, as well as the Si i, shows no difference between observations inside and outside of transit and this is presumably the stable stellar line. Assuming thermal broadening with T = 10⁴ K, we place a p = 0.01 upper limit on any total He i equivalent width as <25 mÅ. There are residuals associated with the two Si i and Na i lines, i.e., a shift and increase in depth of ∼1% during the transit. This cannot be due to the planet (transit depth of 0.16%) and is unlikely to be due to the motion of spots on the rotating star because the overall variability in the K2 lightcurve is only 0.3% and rotation is comparatively slow (15 days). Such an effect could, in principle, be produced if the transit was actually produced by an eclipsing companion star with a much different effective temperature (see below), but this scenario is ruled out by the shape of the K2 lightcurve and the lack of secondary eclipses (Ciardi et al. 2018; Mann et al. 2018). The line at 1083.34 nm is presumably stellar but not in the NIST Atomic Line Database.

Zoom In Zoom Out Reset image size

The absence of detectable He i may be due to the properties of the planet and/or low irradiation by the host star, responsible for both driving escape and ionizing He which recombines into the neutral triplet state (Oklopčić & Hirata 2018). The mass of K2-136c is not known but a high mass and metal-rich atmosphere could inhibit escape. K2-136 is neither rapidly rotating nor exceptionally active; it has a rotation period of 15 days and a mass of 0.74M_⊙ (Mann et al. 2018). For a Hyades age of 650 Myr (Martín et al. 2018) the predicted convective turnover time is 36 days (Dotter et al. 2008) and the Rossby number is 0.4, i.e., an unsaturated dynamo. The empirical relation of Wright et al. (2018) predicts L_X/L_BOL ≈ 8 × 10⁻⁵ which means an expected L_x  = 1.3 × 10⁵ L_⊙ or an X-ray flux of 1.2 × 10⁻¹³ erg s⁻¹ cm⁻². This is below the detection limit of ROSAT near the ecliptic plane (Boller et al. 2016), which explains why it does not appear in the 2RXS source catalog. The predicted X-ray irradiance of K2-136c is 1250 erg s⁻¹ cm⁻². This is below estimates for other planets around young stars (AU Mic, K2-25b, K2-100b) where He i was not detected (Gaidos et al. 2020a, 2020b; Hirano et al. 2020a). K2-136 does appear as a near-UV source in the Galex GR6+7 AIS catalog (Bianchi et al. 2017), but its V − J = 2.10 and NUV − K_s  = 11.10 place it among inactive stars in a color–color diagram (Ansdell et al. 2015). Curiously, K2-136 has a candidate late M-type stellar companion at a projected separation of 40 au (Ciardi et al. 2018); stars with such companions are typically more rapidly rotating and active (Stauffer et al. 2018; Messina 2019) than their single counterparts, but K2-136 seems to be an exception to this rule.