The Two Hot Corinos of the SVS13-A Protostellar Binary System: Counterposed Siblings

Figure 1 shows the SVS13-A map of the dust continuum emission at 1.3 mm. The two emission peaks correspond to the two components of the SVS13-A system previously observed with ALMA as well as at centimeter wavelengths (Carrasco-González et al. 2008; Tobin et al. 2018; Diaz-Rodriguez et al. 2021): VLA4A and VLA4B. In addition, the map shows arc-like structures in agreement with that observed by Tobin et al. (2016, 2018) and Diaz-Rodriguez et al. (2021): a bright arc (up to 32σ) extending from the southern edge of the continuum associated with VLA4A toward the southeast and two fainter arcs (up to 16σ) parallel to the first one and extending on the east (i.e., from the southern edge of the continuum associated with VLA4B) and on the west side of VLA4A. In addition, the present image shows further three elongated structures extending from the northern edge of the continuum associated with VLA4A and B toward the north, with a slightly negative PA. As the ALMA beam is roughly elongated along the N–S direction (PA = +5°), we checked for the presence of possible artifacts caused by the cleaning process by repeating the cleaning using different tapering. With a larger circular beam, all the elongated arc-like structures, including the three structures toward the north, are less evident but still present. The presence of accretion streamers, suggested for VLA4B (Diaz-Rodriguez et al. 2021), is then confirmed by the present ALMA data set also toward VLA4A, and it has an impact on the following discussion on the origin of iCOM emission (Section 5.1).

Zoom In Zoom Out Reset image size

The continuum emission associated with VLA4A and VLA4B has a disk-like structure. We modeled the continuum emission in the uv plane assuming two elliptical inclined disks centered on the continuum peak positions. The best fitting is obtained with one disk with axes 0602 (0001) × 0231 (0077) and PA 41.1° (01) centered toward VLA4A and a second disk with axes 0325 (0242) × 0205 (0153) and PA 338 (01) centered toward VLA4B. The integrated fluxes are 183.15 (0.87) mJy and 126.06 (0.55) mJy for VLA4A and VLA4B, respectively. The sizes are ∼180 × 70 au (VLA4A) and 90 × 70 au (VLA4B). From the axis ratio, we can infer an inclination of 67° and 51° (with respect to the plane of the sky) for the VLA4A and VLA4B disks, respectively. We calculate the disk masses assuming isothermal and optically thin dust emission. If we assume a dust temperature of 140 K, as derived from line analysis, a dust-to-gas ratio of 1/100, and a dust mass opacity coefficient k_ν of 1.0 cm² g⁻¹ (Ossenkopf & Henning 1994), we obtain 0.029 M_⊙ for VLA4A and 0.020 M_⊙ for VLA4B. The value obtained for VLA4B is consistent with that reported by Diaz-Rodriguez et al. (2021), while the one for VLA4A is higher by a factor of 3. If we assume a lower dust temperature of 30 K, we obtain 0.159 M_⊙ for VLA4A and 0.110 M_⊙ for VLA4B, in good agreement with that found by Tobin et al. (2018).

We detected a total of 17 emission lines from 4 iCOMs (Table 1). More specifically, we detected four lines of CH₃OH and five lines of ¹³CH₃OH, covering upper-level energies (E_up) from 57 K to 804 K. We also detected two CH₃CHO lines (84 K), four lines of CH₃OCH₃ (26–226 K), and three lines of NH₂CHO (79–127 K). In Figure 2, we show the emission of one representative line for each species, integrated on three velocity intervals, namely the CH₃OH 25_3,22–25_2,23 A at 241.589 GHz; the ¹³CH₃OH 14_3,12–14_2,13 A line at 256.827 GHz; the CH₃CHO 13_1,13–12_1,12 E line at 242.106 GHz; the 13_1,13–12_0,12 EA, AE, EE, and AA transitions at 241.946 GHz; and the NH₂CHO 12_4,9–11_4,8 transition at 255.059 GHz. The lines are representative of each species, independently of their upper-level energy. All the maps are reported in Figures C1–C5. Figure 2 also shows the iCOM spectra extracted in the positions corresponding to the continuum peak position of the two sources VLA4A and VLA4B. From the ¹³CH₃OH spectra, we reveal different systemic velocities for VLA4A (+7.7 km s⁻¹) and VLA4B (+8.5 km s⁻¹). The spatial distribution of iCOMs is different for the different species. For CH₃OH and ¹³CH₃OH, in the interval between +4.0 and +7.5 km s⁻¹, the line emission traces VLA4A, with an elongation toward the north. In the velocity intervals 7.5–8.7 km s⁻¹, corresponding to the systemic velocity reported for SVS13-A, and 8.8–12.0 km s⁻¹, the line emission is associated with both protostars, peaking in between. A beam deconvolved size of 06 × 04 is derived from a simple 2D Gaussian fit in the image plane of the ¹³CH₃OH moment 0 emission. For CH₃OCH₃, the spatial distribution is similar but the emission between +8.8 and +11 km s⁻¹ peaks closer to VLA4B. Acetaldehyde emission between 8.8 and 11 km s⁻¹ peaks closer to VLA4A. Finally, formamide, which is the only N-bearing species detected in our observations, is peaking toward VLA4A.

Zoom In Zoom Out Reset image size

In Table 1 we report the integrated line intensities extracted from the continuum peak positions of VLA4A and VLA4B, respectively, and their intensity ratios. The intensity ratios are close to the value of 1 for CH₃OCH₃, confirming that dimethyl ether is present in both protostars. For methanol and acetaldehyde, intensity ratios vary from 1.3 to 3.2, while formamide abundances are higher in VLA4A by factors of 4–7. The intensity ratios suggest chemical differentiation inside the system. To further investigate the species spatial distribution, we compare emission lines with similar E_up (between 57 and 81 K) in order to minimize excitation conditions effects. Figure 3 shows the normalized integrated line profiles of CH₃OCH₃ and NH₂CHO, extracted along the horizontal axes connecting the two protostars (Figure 2). The x-axis reports the offset in astronomical units with respect to the position of VLA4A. Two vertical dashed lines indicate the positions of the two protostars. The green line is the residual emission obtained subtracting the NH₂CHO spectra from the CH₃OCH₃ emission. If we assume that NH₂CHO mainly traces VLA4A, the green profile is the additional emission from VLA4B. Figure 3 also shows the residual profiles of CH₃OCH₃, CH₃OH, ¹³CH₃OH, and CH₃CHO after subtracting the NH₂CHO emission. The figure highlights the emission of different species at the position of VLA4B. Finally, Figure A1 shows the intensity-weighted peak velocity (moment 1) maps of the brightest ¹³CH₃OH, CH₃OCHO, and NH₂CHO lines. All of the lines show velocity gradients roughly orthogonal to the outflow direction. Methanol shows the broadest velocity interval while formamide shows the tightest.

Zoom In Zoom Out Reset image size

In order to characterize the two hot corinos, we analyzed their methanol emission on the dust peaks (Table 1) via the non-LTE LVG (large velocity gradient) code grelvg, developed by Ceccarelli et al. (2003). We used the collisional coefficients with para-H₂, computed by Rabli & Flower (2010) between 10 and 200 K for the first 256 levels and provided by the BASECOL database (Dubernet et al. 2013). We assumed the A/E CH₃OH ratio equal to 1 (Flower et al. 2006) and ¹²C/¹³C = 60 (e.g., Milam et al. 2005). To compute the line escape probability as a function of the line optical depth, we adopted a semi-infinite slab geometry and a line width equal to 4 km s⁻¹, as measured. We ran a large grid of models (∼10,000) to cover the χ² surface in the parameters space: the total (A- plus E- ) methanol column density N(CH₃OH) from 1 × 10¹³ to 2 × 10¹⁸ cm⁻², the H₂ density n_H2 from 1 × 10⁵ to × 10¹⁰ cm⁻³ and the temperature T from 40 to 200 K. We fitted the measured ¹³CH₃OH-A line intensities by comparing them with those predicted by the model, leaving N(¹³CH₃OH), n_H2, and T as free parameters. We obtained a good fit (χ² = 1.1) with the following parameters: N(¹³CH₃OH) = (1.0 ± 0.2) × 10¹⁷ cm⁻², T = (140 ± 40) K, and n_H2 = 3 × 10⁸ cm⁻³ with a lower limit of 2 × 10⁷ cm⁻³, indicating that the lines are LTE populated. ⁶ Line optical depths are estimated using the non-LTE predictions. The predicted opacity of the ¹³CH₃OH-A line is between 0.1 to 0.4. Because the ¹²CH₃OH-E lines are predicted to be very optically thick (τ ∼ 60), we derive the column density of the main isotopologue from the ¹³CH₃OH one, assuming the same temperature of 140 K, and we obtain N(CH₃OH) = (6 ± 1) ×10¹⁸ cm⁻². We repeated the same analysis also for VLA4B and obtained the following parameters (reduced χ² = 0.95): N(CH₃OH) = (5 ± 1) × 10¹⁸ cm⁻², T = (170 ± 50) K, and n_H2 = 3 × 10⁸ cm⁻³ with a lower limit of 2 × 10⁷ cm⁻³, again indicating that the lines are LTE populated. Please note that the line with E_up = 800 K cannot be modeled by the non-LTE code because of the lack of collisional coefficients. Assuming that the line upper level is LTE populated, the predicted intensity is about a factor of 2 lower than the observed one. ⁷ Yet, it is very likely this high-lying line is subthermally populated so that the predicted intensity even has to be considered an upper limit. It has already been found by other authors that high-lying (with E_up ≥ 500 K) methanol lines are most likely due to radiatively populated methanol rather than collisionally populated, as discussed in Leurini et al. (2007).

Because the methanol non-LTE analysis showed that both the VLA4A and VLA4B hot corinos have large densities, we carried out an LTE rotational diagram analysis for the other iCOMs, assuming a rotational temperature of 140 and 170 K for VLA4A and VLA4B, respectively, as derived from the methanol analysis. The results of the analysis are reported in Table 2. Synthetic LTE spectra generated using Weeds (Maret et al. 2011) are reported in Figures E1–E4. The column density ratios between VLA4A and VLA4B are consistent with the line intensity ratios reported in Table 1. In Table 2 we also report the column density ratios of the other iCOMs with respect to methanol, which is in agreement with the ASAI analysis (Bianchi et al. 2019).

Table 2. Results of the Non-LTE LVG and LTE Rotational Diagram Analysis of the iCOM Emission Observed toward VLA4A and VLA4B

	VLA4A			VLA4B			VLA4A/VLA4B a
Species	Trot	Ntot	X ${}_{{\mathrm{CH}}_{3}\mathrm{OH}}$ b	Trot	Ntot	${X}_{{\mathrm{CH}}_{3}\mathrm{OH}}$ c
	(K)	(cm−2)		(K)	(cm−2)
Non-LTE Analysis
CH3OH c	140 d	6(1) × 1018	...	170 d	5(1) × 1018	...	1.2 (0.3)
13CH3OH	140(40)	1.0(0.2) × 1017	...	170(50)	8(2) × 1016	...	1.3 (0.4)
Rotational Diagram Analysis
CH3CHO	140 d	2.4(0.4) × 1016	4.0 (1.0) × 10−3	170 d	1.1(0.2) × 1016	2.2 (0.6) × 10−3	2.2 (0.5)
CH3OCH3	140 d	7.5(1.2) × 1016	1.3 (0.3) × 10−2	170 d	1.0(0.2) × 1017	2.0 (0.6) × 10−3	0.8 (0.2)
NH2CHO	140 d	5.9(0.9) × 1015	4.0 (0.9) × 10−3	170 d	1.4(0.3) × 1015	2.8 (0.8) × 10−4	4 (1)

Notes.

^a Column density ratios. ^b Abundance ratio with respect to CH₃OH. ^c Derived from ¹³CH₃OH assuming ¹²C/¹³C = 60. ^d Assumed, as derived by the methanol non-LTE analysis.

Download table as:  ASCIITypeset image

The three continuum arc structures observed toward the south and the three fingers toward the north could be either accretion streamers or alternatively, outflow cavity walls. The brightest arc-like structure in the south has been reported on similar spatial scales by Tobin et al. (2018) in the 1.3 mm continuum emission and in blueshifted C¹⁸O(2–1) emission. Tobin et al. (2018) stressed that this one-armed spiral pattern is similar to that observed in L1448 IRAS3B (Tobin et al. 2016) and associated with a disk inclined by ∼45° surrounding the triple system. The SVS13-A disks are closer to edge on (i = 67° and 50°) than L1448 IRAS3B. Although the contribution of cavities opened by the precessing jets/outflows cannot be excluded, recent studies have shown that accretion streamers feeding disks are common phenomena from Class 0 to Class II objects (Yen et al. 2019; Pineda et al. 2020; Garufi et al. 2022), where planet formation is possibly already ongoing. Their presence may strongly affect the disk chemical composition because they can produce slow shocks when impacting the disk. The chemistry occurring in the shocked gas could synthesize new molecules that would eventually accrete into the disk and, consequently, enrich it. Further observations of typical shock tracers such as SO and SO₂ (e.g., Garufi et al. 2022) are needed to confirm the presence of accretion streamers in VLA4A and VLA4B.

On the other hand, iCOM emission is compact, confirming the presence of two hot corinos, one in VLA4A and a second one in VLA4B. The molecular emission sizes (03–05) are consistent with thermal sublimation of the icy mantles, namely the classical definition of a hot corino (Ceccarelli 2004). The velocity gradient roughly perpendicular to the outflow direction (see Figure A1) suggests the presence of rotating gas, either from two protostellar disks or from the inner portion of the infalling envelope.

Thanks to the superb angular resolution (013 in the east–west direction), the blueshifted line emission appears to be elongated in the north direction up to ∼150 au, being spatially coincident with the northern accretion streamers (Figure C1–C5). The iCOM emission could be associated with material accreting from the streamer and inducing a slow shock in the protostellar disk, similarly to what was found in, e.g., L1527 (Sakai et al. 2014; Oya et al. 2015 ; Sakai et al. 2017). Moreover, we cannot exclude the possibility that the system has a higher multiplicity, including multiple unresolved cores. Observations of different molecular tracers at extremely high angular resolution (tens of astronomical units), as well as an accurate physical model of the system, are necessary to conclude the origin of iCOM emission.

Finally, the data suggest chemical segregation between formamide, tracing mainly VLA4A and the rest of the observed iCOMs, in the inner 100 au of the system. In Appendix D we discuss the possibility that the observed chemical segregation is due to opacity effects. While dust and line opacity effects cannot completely be excluded, our data suggest a real chemical differentiation of the two hot corinos instead of temperature gradients or excitation effects.

A different spatial distribution between O-bearing and N-bearing molecules has been observed in Orion KL (Blake et al. 1987; Peng et al. 2013, and references therein) and in other high-mass star-forming regions on scales of thousands of astronomical units (Allen et al. 2017; Csengeri et al. 2019). Investigating whether a similar segregation is present also in the inner 100 au around low-mass protostars is more complicated as only a few system have been observed with enough angular resolution. An intriguing chemical segregation between formamide and acetaldehyde has been observed in the jet-induced shock L1157-B1, at a distance of >0.1 pc from the protostar (Codella et al. 2017). In this case, the dichotomy is reproduced well by chemical models that assume that the formamide formation is dominated by gas-phase reactions (chemistry effect). A chemical differentiation has also been observed in the prototypical Class 0 binary system IRAS 16293–2422 (Manigand et al. 2020). The abundance of CH₃OCH₃ relative to methanol is similar in the two components of the IRAS 16293–2422 system, while it differs by a factor of 2 for CH₃CHO and 4 for NH₂CHO, similarly to what was found in VLA4A and VLA4B. In IRAS 16293–2422, the observed differences are interpreted as a result of the onion-like structure of the hot corino (physical effect). In particular, the NH₂CHO rotational temperature is slightly higher (T ∼ 140–300 K) than what was derived for other iCOMs (T ∼ 100 K) (Jørgensen et al. 2016, 2018; Manigand et al. 2020), suggesting that N-bearing species trace hotter gas, closer to the protostar. On the other hand, recent ALMA observations of other two Class 0 protostars, Perseus B1-c and Serpens S68N, do not show a significant difference in the excitations conditions of N-bearing and O-bearing species (van Gelder et al. 2020; Nazari et al. 2021), leaving the question open. Further observations in different N-bearing and O-bearing complex species are required to clarify the origin of the chemical segregation observed in SVS13-A and inferred in IRAS 16293–2422 A and B but not in other low-mass protostars. The spatial distribution of iCOM emission in SVS13-A supports the idea that formamide is formed in a compact region closer to the protostar and thus, contrarily to other iCOMs, can be more easily obscured by the dust. The fact that the dust emission of VLA4B is optically thicker than that of VLA4A supports this scenario. Note also that the binding energy, recently computed by Ferrero et al. (2020), of formamide (average ∼8400 K) is larger than the one of methanol (average ∼6200 K), in agreement with the possible onion-like structure of the two SVS13-A hot corinos. Interestingly, Diaz-Rodriguez et al. (2021) report the detection of ethylene glycol only toward VLA4A, suggesting that the observed chemical segregation in SVS13-A is not exclusive of formamide. Because no binding energy of ethylene glycol is available in the literature, we carried out new ab initio quantum chemistry calculations to evaluate its binding energy over an 18 water molecule cluster (see Appendix F). In order to compare ethylene glycol with formamide, we also recalculated the binding energy of the latter. We found that the binding energy of ethylene glycol has an average value of ∼7100 K, while that of formamide is equal to ∼4700 K. We emphasize that the difference with respect to Ferrero et al. (2020) is due to a different water cluster used. The important point here is the relative binding energy of ethylene glycol with respect to formamide, which is not expected to change with the water cluster used. This means that if the calculation by Ferrero et al. (2020) was performed for ethylene glycol, the binding energy would be even higher than that of formamide (∼8400 K). This justifies the similar spatial distribution of formamide and ethylene glycol, different from that of methanol. Overall, the new calculated binding energies support the hypothesis that the observed chemical differentiation is caused by the onion-like structure of the two SVS13-A hot corinos instead of spatial segregation between O-bearing and N-bearing species. Further observations at high angular resolution (tens of astronomical units) with ALMA and at centimeter wavelengths with JVLA (and, in the future, with ngVLA and SKA) are needed to further investigate the chemical stratification inside the two hot corinos VLA4A and VLA4B and to quantify the dust opacity effects.