Resolving the Binary Components of the Outbursting Protostar HBC 494 with ALMA

We started our analysis by fixing the visibilities’ phase center on J2000 05h40m27.448s -07d27m29.65s. The image synthesis of the 1.3 mm continuum emission was performed with the tclean task of CASA. For the high-resolution data, we used the Briggs weighting scheme with a robust parameter of 0.5, resulting in a synthesized beam size of 41.2 $\times$ 29.8 mas, and a position angle of 40.6. The pixel scale of the image was set to 1.5 mas. One iteration step of self-calibration was applied to both observations. For the first night (low-resolution image: 142 mas), it resulted in an improvement of $\sim$6 SNR. For the second night (high-resolution image: 27 mas), it resulted in an SNR improvement of $\sim$ 1.6.

Our final continuum image was produced from a concatenated measurement set combining both, the short and the long baseline visibilities, using the concat task of CASA. We used the same image synthesis parameters used for the second night of observation.

The visibilities were fixed at the same phase center as was done for the continuum coordinates (see subsection 2.1). After producing a dirty image from the visibilities, we noticed that the emission mimicked the presence of gaseous envelopes around the disks (moment 0 maps, see Figure 1, left column) but no gas dynamic signatures were revealed (moment 1 map, see Figure 1, right column). Therefore, we proceeded to remove the visibility’s continuum contributions by using the CASA task uvcontsub.

Next, we performed the tclean process using Briggs weighting with a robust parameter of 0.5, Hogbom deconvolver, and pixel scale of 6 mas. The total velocity ranged between -6 km/s to 18 km/s, although a smaller range (0 km/s to 11 km/s) was used to create the ¹³CO and C¹⁸O moment maps. We used spectral resolution widths of 1 km/s for ¹²CO and of 0.3 km/s for ¹³CO and C¹⁸O. The rms values, obtained for the channel without a clear signal on the channel maps (4 km/s), were 1.9, 3.6, and 2.6 mJy/beam for ¹²CO, ¹³CO and C¹⁸O, respectively.

With the high-resolution data shown in Figure 2, we could reveal the binary nature of the HBC 494 system. Each of the components in HBC 494 is surrounded by continuum emission, most likely associated with circumstellar disks.

A large fraction of the mm-sized grains in protoplanetary disks resides in the midplane. They are usually optically thin to their own radiation. Thus, assuming standard disks, we can use the observed fluxes to estimate the dust masses of HBC 494 N and S. For that, we use the following formula (as in Beckwith et al., 1990; Andrews & Williams, 2005; Cieza et al., 2018):

The next two subsections describe the large and small-scale structure analyses. The large scale, using the extent of 8000 au, shows the results of outflows and envelopes. At such a scale, there is not enough resolution to display both disks’ dynamics and their possible interactions. The small scale (150 au) fulfills this role, and thus, both gas scenarios must be taken into account.

Besides HBC 494, there are $\sim$ 25 known FUor objects within 1 kpc (Audard et al., 2014). As previously commented, many triggering mechanisms can cause episodic accretion events. From this sample, only HBC 494 and a few other FUor systems are known binaries (e.g., FU Orionis, L1551 IRS, RNO 1B/C, AR 6A/B; Pueyo et al., 2012 and references therein).

Given the high occurrence rate of stellar binaries harboring disks, the lack of detection of multiple YSO systems remains a case of investigation. Additionally, young disks evolve in crowded star-forming regions, enhancing the hypothesis that the outbursts of Class 0/I disks may affect other disks, driving episodic accretion events. Also, surveys and studies using ALMA and NIR data for Class I-III disks, show that there is a lack of detection when only visual multiple systems (separations of 20-4800 au) are considered (e.g., Zurlo et al., 2020a, 2021). When all the possible separations are taken into account, the multiplicity frequencies considerably increase. For Taurus, for example, it goes up to 70% when spectroscopic binaries are included (Kraus et al., 2011). In the case of Orion, 30% of the systems are multiple with separations between 20 and, 10000 au (Tobin et al., 2022). The latter study also noticed that the separations decrease with time (when comparing Class 0, I, and flat-spectrum disks) and that the multiplicity frequency in Class 0 is higher than in the later evolutionary stages. Therefore, multiple eruptive young systems must be common, and more detectable within evolutionary time. They probably are not frequently detected due to the short time duration of enhanced accretion events in comparison with the lifetime of disks in class 0/I stages (see Audard et al. 2014 for a FUors review). Still, it is not clear if disks belonging to close-separation systems are more susceptible to eruptive events than isolated ones.

For each HBC 494 disk, no clue of trigger mechanisms was found, since the gas and continuum data did not reveal clear spiral or clumpy features indicating a case of infall or GI (see e.g., Zhu et al., 2012; Kratter & Lodato, 2016). Looking at the dust continuum (Figure 2), CO isotopologues moment maps (Figure 7), and the scattered ¹²CO emission (Figure 12 of Ruíz-Rodríguez et al. 2017), the hypothesis of stellar flybys is also not encouraged since no trace of perturbation was detected (Cuello et al. 2020, 2023). Moreover, dynamical data (small-scale moment 1 maps) also can provide clues for the triggering mechanism (Vorobyov et al., 2021). However, it requires clear detection of quasi-keplerian rotation, which was not observed.

In the case of binaries, the first component to ignite the FUor outbursts can quickly trigger the secondary one by inducing perturbations and mutual gravitational interactions (Bonnell & Bastien, 1992; Reipurth & Aspin, 2004; Vorobyov et al., 2021; Borchert et al., 2022a, c). Therefore, we can assume that all the disks in FUor multiple systems like HBC 494 might have undergone successive enhanced eruptive stages, despite the differences in mass and radius between each component. However, it is not clear how the outbursts in HBC 494 N and HBC 494 S affect each other since the moment maps do not show the connection between disks. Although, on larger scales, we detected the difference in the amount of gas mass detected in the northern regions compared to the southern. This scenario may be described if the more massive disk (i.e. HBC 494 N) has a higher contribution to the outbursts and winds.

With astrometric data, the eccentricity of orbits can be determined. If both disks are still accreting gas from the environment, it is expected that quasi-circular orbits trigger enhanced accretion every few binary periods. Eccentric orbits, on the other hand, can induce eruptive events in every orbit, preferentially when the binaries reach the pericentre (see e.g. Dunhill et al., 2015; Lai & Muñoz, 2022). Observational evidence of circumbinary disks in different YSO stages was found (see e.g. Dutrey et al., 1994; Mathieu et al., 1997; Tobin et al., 2016a; Maureira et al., 2020) and are seen as a common counterpart of young binaries in formation. However, no HBC 494 circumbinary gas disk was observed due to cloud contamination and optical depth effects.

FU Orionis, the precursor of the classification FUor, is also a binary system. HBC 494 system, however, has smaller disks separation compared to the 210 au between FU Orionis north and south components (Pérez et al., 2020). FU Orionis components have similar sizes and masses between their disks, contrary to the HBC 494 components. We can argue that, due to HBC 494 being a close-packed system with substantially different masses between their components, a scenario of radius truncation can be tested. However, the FU Orionis disks are exceptional compared to other FUors, which are generally more massive and larger. Nevertheless, it is important to check if HBC 494 follows the trend which shows that FUor objects are more massive than class 0/I disks and that disks from multiple systems are smaller compared to those from non-multiple systems (Cieza et al., 2018; Hales et al., 2020; Tobin et al., 2020; Zurlo et al., 2020b). First, we will describe our results regarding masses and radius. Then, we will compare HBC 494 disks to FUors and Orion YSOs in the literature.

We assumed dust masses (assuming the gas-to-dust ratio of 100) inferred by optically thin approximation (described in section 3.1). However, it is worth mentioning that the optically thin analysis can not take into account the innermost regions of the disk (optically thick). This may lead to an underestimation of the dust masses. Also, an underlying miscalculated dust grain temperature can alter the results, overestimating the masses if the dust is warmer than expected. The chosen assumptions of temperatures (fixed 20 K for dust grains) and opacities for both disks lead to 1.43 $M_{\rm Jup}$ and 0.29 $M_{\rm Jup}$ for HBC 494 N and HBC 494 S, respectively. Based on these results, we can say that the HBC 494 N is comparable to other FUor sources, but HBC 494 S has its dust mass comparable to EXor disks (Cieza et al., 2018). The disk sizes were calculated using the deconvolved Gaussian FWHM/2 radius obtained from 2D Gaussian fits. This methodology was also used for other ALMA datasets (e.g., Hales et al., 2020; Tobin et al., 2020). Therefore, it allows a consistent comparison with other disks. The dust mass vs radius comparison between HBC 494 N and HBC 494 S, a sample of 14 resolved FUors and EXors (extracted from Hales et al., 2020) and Class 0/I systems (Tobin et al., 2020), can be seen in figure 8. Here, as previously stated, we notice that HBC 494 disks (black, hexagon symbols), as other eruptive systems (hexagon, square, and triangle symbols), present higher masses than young systems, which are not in the stage of episodic accretion. However, there is no obvious distinction between the disk sizes of individual systems, eruptive or not. A clear difference is seen when we compare the disks from multiple systems (hexagon symbols) and single systems. The latter present bigger sizes as they are not targets for tidal truncation, enhanced radial drift, and more aggressive photoevaporation — processes known to rule close-systems dynamics and evolution (see e.g., Kraus et al., 2012; Harris et al., 2012; Rosotti & Clarke, 2018; Zurlo et al., 2020a, 2021).

Different scenarios have been proposed to explain the formation of binary systems. The main one for close-separated systems is fragmentation, followed by dynamic interactions (for a review, see Offner et al., 2022).

The fragmentation process consists of partial gravitational collapse from self-gravitating objects. To be successful, many initial physical conditions are relevant as thermal pressure, density, turbulence, and magnetic fields. In addition, the fragmentation can be divided into two main classes: direct/turbulent (e.g., Boss & Bodenheimer, 1979; Bate & Burkert, 1997) and rotational (e.g., Larson, 1972; Bonnell, 1994; Bonnell & Bate, 1994a, b; Burkert et al., 1997).

The direct/turbulent fragmentation from a collapsing core is highly dependent on the initial density distribution. It can form wide-separated multiple YSO systems with uncorrelated angular momentum. Thus, the direct/turbulent fragmentation may lead to preferentially misaligned multiple systems (Bate, 2000; Offner et al., 2016; Bate, 2018; Lee et al., 2019). The rotational fragmentation, instead, is caused by instabilities in rotating disks and leads to preferentially spin-aligned and coplanar systems (Offner et al., 2016; Bate, 2018). Moreover, later dynamical processes such as stellar flybys (Clarke & Pringle, 1993; Nealon et al., 2020; Cuello et al., 2023) and misaligned accretion from the environment (Dullemond et al., 2019; Kuffmeier et al., 2020, 2021) can also induce alignment or misalignment.

The HBC 494 disks have similar inclinations ($\Delta i$ = 7.5 $\pm$ 4.1 degrees) and similar PAs ($\Delta PA$ = 4.6 $\pm$ 7.0 degrees). The relative orientations of the disks suggest that the system is coplanar. Assuming they are quasi-coplanar, with the same PA, the unprojected separation would be $\sim$0$\aas@@fstack{\prime\prime}$24 (99 au). Therefore, the HBC 494 disks may show a spin-alignment situation, more typical to observed close-separated systems (see, e.g., Tobin et al., 2016b, 2020). Thus, we tentatively suggest that HBC 494 was formed by rotational disk fragmentation rather than direct collapsing. However, more precise observation of the kinematics of the surrounding gas on a small scale (Vorobyov et al., 2021), complemented with orbital information coming from astrometrical measurements, is required to evaluate this hypothesis.

In this context, IRAS 04158+2805 is of interest as Ragusa et al. (2021) reported the detection of two circumstellar disks, similar to HBC 494 in terms of small to moderate misalignment, and an external circumbinary disk. However, in contrast to IRAS 04158+2805, no circumbinary disk was detected in HBC 494. This difference could be explained either by the absorption of the circumbinary disk emission due to cloud contamination, or because HBC 494 had enough time to sufficiently empty the circumbinary mass reservoir. Since most hydrodynamical models of star formation naturally produce young binaries with individual disks and a surrounding circumbinary disk (Bate, 2018, 2019; Kuruwita et al., 2020), it is likely that a circumbinary disk formed at some earlier evolutionary stage of HBC 494. If the circumbinary disk is actually there, but remains undetected, the supply of material could extend the disk lifetime of the circumstellar disks in HBC 494. Deeper multi-wavelength observations would provide more information, enabling us to solve this binary riddle.