A CANDIDATE PLANETARY-MASS OBJECT WITH A PHOTOEVAPORATING DISK IN ORION

We observed the central object of Proplyd 133-353 during the night of 2015 December 31 in the MOS mode with MMIRS, as part of our spectroscopic survey of young stars in Orion. We used the HK grism and the HK3 filter and created a mask with 05 × 7'' slits, following the MMT mask preparation procedure. This setup yields spectra from 1.25 to 2.34 μm, with an average spectral resolution of λ/Δλ ≈ 1100. We took 8 exposures (8 × 300 s) for 4 spatially dithering pairs. After the scientific exposures, we observed one telluric standard HD 34481 at a similar airmass (airmass differences <0.02) to our scientific targets.

We reduced the data using the CFA MMIRS pipeline from Chilingarian et al. (2015). The pipeline can reduce the MMIRS spectroscopic data in a fully automatic way. It subtracts pairs of dithered spectral exposures, performs flat-field and residual sky corrections, and rectifies the 2D spectra. The spectra are wavelength calibrated using airglow OH lines and corrected for the telluric absorption using telluric standard stars we observed. A detailed description of the pipeline can be found in Chilingarian et al. (2015).

We constructed a set of spectral templates with spectral types ranging from G5 to M9.5, using the spectra of diskless young stars collected with the instrument X-Shooter mounted on the Very Large Telescope (Manara et al. 2013, 2016, in preparation). Taking the visual extinction as a free parameter, we fit the observed spectra using the spectral templates that are reddened with the extinction law from Cardelli et al. (1989), adopting a total to selective extinction value R_V = 3.1. The best fit is determined by minimizing ${\chi }_{\mathrm{red}}^{2}=\tfrac{1}{N-1}{\sum }_{i=1}^{N}\tfrac{{({F}_{\lambda ,\mathrm{obs}}-{F}_{\lambda ,\mathrm{template}})}^{2}}{{\sigma }_{\lambda ,\mathrm{obs}}^{2}}$, where N are the numbers of the wavelength bins, ${F}_{\lambda ,\mathrm{obs}}$ and ${F}_{\lambda ,\mathrm{template}}$ are the observed fluxes and the ones from the spectral templates, respectively, within the wavelength λ bin, and ${\sigma }_{\lambda ,\mathrm{obs}}$ are the errors in the observed fluxes. In Figure 2, we show a comparison of the observed spectrum of Proplyd 133-353 and the X-Shooter templates reddened with the best-fit visual extinction corresponding to the individual templates. We note that the template with spectral type M9.5 gives the best fit to the observation. However, a conversion of the spectral type M9.5 to the effective temperature (T_eff) is still uncertain and could range from less than 2400 K to ∼2500 K using the different temperature scales in the literature (Luhman et al. 2003; Rajpurohit et al. 2013; Herczeg & Hillenbrand 2014). To avoid this uncertainty, we fit the spectrum of Proplyd 133-353 directly with the BT-Settl atmospheric models with solar abundances from Asplund et al. (2009) and a surface gravity Log g = 3.5, which is suitable for the very low-mass young substellar objects (Baraffe et al. 2015). As with our procedure for the X-Shooter templates, we took the visual extinction as a free parameter and searched for the best-fit atmospheric models. Figure 2 shows a comparison of the observed spectrum and the reddened BT-Settl models. We found the best-fit BT-Settl models are the ones with the effective temperatures 2400 and 2500 K with the best-fit visual extinction of 0.3 and 0.7 mag, respectively. Thus, in this work, we adopt T_eff = 2450 ± 50 K and A_V = 0.5 ± 0.2 for the central object of Proplyd 133-353.

Zoom In Zoom Out Reset image size

We obtained the near-infrared photometry of Proplyd 133-353 from Robberto et al. (2010). The stellar luminosity of Proplyd 133-353 was derived using dereddened J-band photometry, and the J-band bolometric correction (BC_J = 2.12) for T_eff = 2450 ± 50 K calculated with the BT-Settl models, assuming a distance of 414 pc (Menten et al. 2007). The stellar luminosity (L_⋆) of Proplyd 133-353 is estimated to be 0.021 ± 0.004 L_⊙. The uncertainty of L_⋆ is calculated considering the uncertainty in the J-band magnitude, the T_eff value, and the extinction. In Figure 3 (left), we placed Proplyd 133-353 in the H–R diagram. As a comparison, we also show other young stars in the ONC with data collected from Da Rio et al. (2012), as well as the pre-main-sequence (PMS) evolutionary tracks from Baraffe et al. (2015). In the H–R diagram, Proplyd 133-353 is above the youngest PMS isochrone (∼0.5 Myr) from Baraffe et al. (2015) and has a much cooler T_eff compared with other young stars, but shows an L_⋆ similar to a 0.05 M_⊙ brown dwarf at 1 Myr. Recently, Kounkel et al. (2016) found the mean distance of the ONC to be 388 ± 5 pc. Using this distance estimate, the L_⋆ of Proplyd 133-353 will be reduced by a factor of 1.14, which moves the object down on the H–R diagram very slightly. Therefore, the object is still very young (age < 0.5 Myr) even with this new distance estimate.

Zoom In Zoom Out Reset image size

Due to the lack of PMS evolutionary tracks at ages younger than 0.5 Myr in Baraffe et al. (2015), it is not possible to derive its mass and age directly through comparison to the PMS evolutionary tracks. Here, we tentatively constrain its mass using the 0.5 Myr isochrone from Baraffe et al. (2015). On this isochrone, the mass of an object with T_eff = 2450 is ∼13 Jupiter mass (M_J), which is much lower than the known lowest mass (42 M_J) for the objects at the centers of proplyds in ONC in the literature (Robberto et al. 2008). However, this mass might be an upper limit for Proplyd 133-353 since the T_eff values of young low-mass substellar objects seem to decrease during their evolution. The dividing line between a brown dwarf and a planet is around 13 M_J, as defined by the deuterium fusion mass limit (Burrows et al. 1997; Spiegel et al. 2011). Therefore, Proplyd 133-353 could be a planetary-mass object with an age younger than 0.5 Myr. We also use the PMS evolutionary tracks from D'Antona & Mazzitelli (1997), which give a mass ∼25 M_J and an age ∼4000 year. We note that evolutionary models at these young ages and very low masses have significant uncertainties (Close et al. 2005; Konopacky et al. 2010); therefore, in this work, we do not attempt to conclusively determine the nature of the object. Whether in the brown dwarf or planetary-mass category, Proplyd 133-353 is an interesting object for a detailed study.

As described in Section 2, Proplyd 133-353 has a tail pointing away from θ¹ Ori C, which indicates that its disk is being dissipated by the massive star. Thus, Proplyd 133-353 should be a member of the ONC. Here, we will discuss further evidence for the ONC membership of Proplyd 133-353 based on its X-ray properties and proper motion.

Proplyd 133-353 has been detected by the Chandra space telescope in Chandra Orion Ultradeep Project (Getman et al. 2005). In Figure 3 (right), we show its X-ray luminosity (L_X) over the 0.5–8.0 keV band versus its stellar luminosity and compare it with the values from other young stars in the ONC with the L_X from Getman et al. (2005) and L_⋆ from Da Rio et al. (2012). The L_X/L_⋆ ratio of Proplyd 133-353 is similar to those of other young stars with similar stellar luminosity.

According to the PMS evolutionary models (Baraffe et al. 2015), the expected L_⋆ for Proplyd 133-353 at an age of 1 Myr is 0.00307 L_⊙, corresponding to its T_eff. If Proplyd 133-353 is a foreground 1 Myr old planet, the expected distance could be ∼160 pc, corresponding to the intrinsic luminosity and the observed brightness. If that is the case, we would expect that Proplyd 133-353 shows a different proper motion from the ones of other young stars in the ONC. We derive the proper motion from the HST archive images in the F775W band observed at two epochs (2005 April versus 2015 February).⁴ In Figure 4, we show the second-epoch HST image. The image has been calibrated in astrometry to the first-epoch image using the common ONC members. We excluded Proplyd 133-353 in the astrometric calibration; therefore, we can measure the proper motion of the proplyd relative to the ONC. As a comparison, in Figure 4, we also show the positions of the stars in the first-epoch HST image. From the image pair, we obtain a relative proper motion for Proplyd 133-353, of μ_α = 4 ± 1 mas yr⁻¹ and μ_δ = −3 ± 1 mas yr⁻¹. Therefore, there is no significant proper motion of Proplyd 133-353 relative to the ONC. However, we cannot exclude the possibility that Proplyd 133-353 and the ONC have different motions in the line of sight, given the spectral resolution (λ/Δλ ≈ 1100) of our MMIRS data.

Zoom In Zoom Out Reset image size

In conclusion, the arguments presented above, especially the cometary structure with a tail pointing away from θ¹ Ori C and the X-ray luminosity, strongly support that Proplyd 133-353 is a member of the ONC.

In the following, we will compare the different scenarios for the origin and nature of Proplyd 133-353.

Theoretical calculations predict that mass-loss rates from disks due to photoevaporation can be on the order of 10⁻⁷ M_⊙ yr⁻¹ within a distance of 0.2 pc from the ionizing massive star θ¹ Ori C (Störzer & Hollenbach 1999), which is confirmed by spectroscopic observations (Henney & O'Dell 1999). The photoevaporation process can effectively dissipate protoplanetary disks outside the gravitation radii (r_g), where the escape velocities equal the speed of sound, which is in turn determined by the UV heating (Hollenbach et al. 2000, p. 401). With such high mass-loss rates, we would expect a very short disk lifetime for Proplyd 133-353, which is ∼10³ years with an assumption of a typical ratio (∼100) between stellar mass and disk mass.

It is well known that there are many tiny dusty clouds, called "globulettes," around H ii regions (De Marco et al. 2006; Gahm et al. 2007; Grenman & Gahm 2014). The mean mass of the globulettes is around several ×10 M_J, and their mean radius is ∼4000 au (Gahm et al. 2007). These globulettes are undergoing photoevaporation by UV photons from the massive stars in H ii regions. However, the shocks generated by photoionization, as well as the surrounding warm gas, can also exert external pressure on the globulettes and drive their collapse to form brown dwarfs and free-floating planets before they are eroded by photoevaporation (Gahm et al. 2007). Such a second-generation formation of brown dwarfs and free-floating planets has been confirmed by the discovery of dense cores in some of the globulettes (Gahm et al. 2013). Haworth et al. (2015) doubt the possibility that the external pressure in the form of radiation or ram pressure can drive globulettes to form brown dwarfs or free-floating planets, and instead they propose the collapse of the globulettes might be triggered by their collisions with the shells that border the expanding H ii regions. If Proplyd 133-353 was formed in such a globulette near θ¹ Ori C due to either mechanism, we might explain both its younger age (with respect to the one of the ONC) and the lifetime of its disk since its parental cloud can protect it from direct photoevaporation for a timescale of 10⁵–10⁶ years (Gahm et al. 2007; Haworth et al. 2015).

Finally, we explore an alternative scenario for the formation of Proplyd 133-353: instead of a photoevaporating disk it may be an evaporating gaseous globule (EGG). In the vicinity of massive stars, a prestellar core can be eroded by the ionizing radiation from massive stars and may form a free-floating brown dwarf or planetary-mass object (Hester et al. 1996; Whitworth & Zinnecker 2004). The final mass of the formed object can be approximated as $0.010\,{M}_{\odot }\,{\left(\tfrac{{a}_{1}}{0.3\mathrm{km}{{\rm{s}}}^{-1}}\right)}^{6}\,{\left(\tfrac{{\dot{N}}_{\mathrm{Lyc}}}{{10}^{50}{{\rm{s}}}^{-1}}\right)}^{-1/3}\,{\left(\tfrac{{n}_{0}}{{10}^{3}{\mathrm{cm}}^{-3}}\right)}^{-1/3}$ (Whitworth & Zinnecker 2004), where a₁ is the isothermal sound speed in the neutral gas of the core, ${\dot{N}}_{\mathrm{Lyc}}$ is the emission rate of Lyman continuum photons, and n₀ is the number density of protons in the H ii region around the core. In ONC, we take ${\dot{N}}_{\mathrm{Lyc}}\sim 2.6\times {10}^{49}\,{{\rm{s}}}^{-1}$ (Bertoldi & Draine 1996), and n₀ ∼ 5.5 × 10³ around Proplyd 133-353 (Weilbacher et al. 2015). Assuming a₁ ∼ 0.23–0.36 km s⁻¹ for a prestellar core with a gas temperature of 15–35 K, the expected final mass of the object formed in the core near Proplyd 133-353 would be around 2–28 M_J, which is consistent with the mass (13 M_J) of Proplyd 133-353.