Hot Corinos Chemical Diversity: Myth or Reality?

Figure 1 reports the map of the continuum emission at 25 GHz. The two continuum peaks mark the two protostars, whose coordinates ($\alpha ({\rm{J}}2000)={03}^{{\rm{h}}}{29}^{{\rm{m}}}10\buildrel{\rm{s}}\over{.} 536$, $\delta (J2000)\,=31^\circ 13^{\prime} 31\buildrel{\prime\prime}\over{.} 07$ for 4A1, and $\alpha ({\rm{J}}2000)={03}^{{\rm{h}}}{29}^{{\rm{m}}}10\buildrel{\rm{s}}\over{.} 43$, $\delta (J2000)=31^\circ 13^{\prime} 32\buildrel{\prime\prime}\over{.} 1$ for 4A2) are consistent with those derived by Tobin et al. (2016) and López-Sepulcre et al. (2017) with higher angular resolution observations. Since the angular resolution of our observations ($\sim 1^{\prime\prime}$) is smaller than the separation between 4A1 and 4A2 (18), they are clearly disentangled in our images, even if individually unresolved with the current resolution.

At centimeter wavelengths, 4A1 shows a brighter continuum emission (due to dust or free–free) than 4A2. The peak fluxes are (2.1 ± 0.3) mJy beam⁻¹ and (0.47 ± 0.07) mJy beam⁻¹ toward 4A1 and 4A2, respectively. Taking into account the slightly different wavelength (1.05 cm) and angular resolution ($\sim 0\buildrel{\prime\prime}\over{.} 1$), these values are consistent with the ones measured by Tobin et al. (2016): (1.3 ± 0.2) mJy beam⁻¹ for 4A1 and (0.38 ± 0.04) mJy beam⁻¹ for 4A2.

All the targeted methanol lines are detected with a signal-to-noise ratio larger than 3 (Table 1). Their velocity-integrated spatial distribution is shown in Figure 1. The methanol emission peaks exactly toward the 4A1 and 4A2 continuum peaks, and it is well disentangled, even if unresolved at the current angular resolution, around the two protostars.

Figure 2 shows the 10 methanol line spectra, isolated and not contaminated by other species, extracted toward the 4A1 and 4A2 continuum peaks. The lines are slightly brighter toward 4A2 than 4A1, whereas the linewidths are very similar (see also Table 1). We derived the velocity-integrated line intensities for each detected CH₃OH transition using a Gaussian fit, being the profile Gaussian-like. The fit results for both sources, namely the integrated emission ($\int {T}_{b}{dV}$), the linewidth (FWHM), the peak velocities (V_peak), and the rms computed for each spectral window, are reported in Table 1. The velocity peaks are consistent with the systemic velocity of the molecular envelope surrounding IRAS 4A (∼6.7 km s⁻¹; Choi 2001). The linewidths are between 3 and 4 km s⁻¹ in agreement with those found by Taquet et al. (2015) and López-Sepulcre et al. (2017) toward 4A2 at millimeter wavelength.

Zoom In Zoom Out Reset image size

In summary, our new VLA observations show a first clear important result: the detection of methanol emission toward 4A1, the protostar where previous millimeter observations showed no iCOMs emission (López-Sepulcre et al. 2017).

To derive the physical properties of the gas emitting CH₃OH, namely gas temperature, density and methanol column density, we performed a non-LTE analysis using a Large Velocity Gradient (LVG) code (Ceccarelli et al. 2003). CH₃OH can be identified in A- and E-type due to the total spin (I) state of the hydrogen nuclei in the CH₃ group: A-type if the total spin function is symmetric (I = 3/2), E-type if asymmetric (I = 1/2) (Rabli & Flower 2010). We used the collisional coefficients of both types of CH₃OH 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 a spherical geometry to compute the line escape probability (de Jong et al. 1980), the CH₃OH-A/CH₃OH-E ratio equal to 1, the H₂ ortho-to-para ratio equal to 3, and that the levels are populated by collisions and not by the absorption of the dust background photons whose contribution is very likely negligible due to the low values of the CH₃OH Einstein coefficients ${A}_{{ij}}$. Please note that the present LVG analysis only accounts for the line optical depth (to have also the dust τ in the methanol emitting region would require information on the structure of the region that we do not have, as the emission is unresolved).

We ran a large grid of models (≥10,000) covering the frequency of the observed lines, a total (CH₃OH-A plus CH₃OH-E) column density ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}$ from 2 × 10¹⁶ to 8 × 10¹⁹ cm⁻², a gas density ${n}_{{{\rm{H}}}_{2}}$ from 1 × 10⁶ to 2 × 10⁸ cm⁻³, and a temperature T from 80 to 200 K. We then simultaneously fitted the measured CH₃OH-A and CH₃OH-E line intensities via comparison with those simulated by the LVG model, leaving ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}$, ${n}_{{{\rm{H}}}_{2}}$, T, and the emitting size θ as free parameters. Following the observations, we assumed the linewidths equal to 3.5 km s⁻¹ and 3.0 km s⁻¹ for 4A1 and 4A2, respectively, and we included the calibration uncertainty (15%) in the observed intensities.

The best fit is obtained for a total CH₃OH column density ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}=2.8\times {10}^{19}$ cm⁻² with reduced chi-square ${\chi }_{R}^{2}=0.6$ for 4A1 and ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}=1\times {10}^{19}$ cm⁻² with ${\chi }_{R}^{2}=0.1$ for 4A2. All the observed lines are predicted to be optically thick and emitted by a source of $0\buildrel{\prime\prime}\over{.} 22$ for 4A1 and $0\buildrel{\prime\prime}\over{.} 24$ for 4A2 (∼70 au) in diameter. Solutions with ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}\geqslant 1\times {10}^{18}$ cm⁻² for 4A2 and $\geqslant 1\times {10}^{19}$ cm⁻² for 4A1 are within 1σ of confidence level. Increasing the methanol column density, the ${\chi }_{R}^{2}$ decreases until it reaches a constant value for ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}\geqslant 1\times {10}^{19}$ cm⁻² for 4A1 and ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}\geqslant 3\times {10}^{19}$ cm⁻² for 4A2; this is because all the observed lines become optically thick ($\tau \sim 10\mbox{--}30$ for 4A1, $\tau \sim 2\mbox{--}6$ for 4A2) and, consequently, the emission becomes that of a blackbody. The results do not change assuming a linewidth ±0.5 km s⁻¹ with respect to the chosen one.

Figure 3 shows, for both sources, the density-temperature ${\chi }^{2}$ surface of the ${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}$ best fit. The gas temperature is (90–130) K for 4A1 and (120–190) K for 4A2, while for the gas density we obtained a lower limit of 2 × 10⁶ cm⁻³ for 4A1 and $1.5\times {10}^{7}$ cm⁻³ for 4A2, which implies that the levels are LTE populated. The fit results are reported in Table 2. The derived ${n}_{{{\rm{H}}}_{2}}$ and T are consistent with those computed with the model summarized in Su et al. (2019) using our sizes.

Zoom In Zoom Out Reset image size

Table 2.  Top: Best-fit Results and 1σ Confidence Level (range) from the Non-LTE LVG Analysis of the CH₃OH Lines toward 4A1 and 4A2

		IRAS 4A1		IRAS 4A2
		LVG Results
		Best Fit	Range	Best Fit	Range
n(H2)	(cm−3)	5 × 106	≥2 × 106	1 × 108	≥1 × 107
${T}_{\mathrm{kin}}$	(K)	100	90–130	160	120–190
${N}_{{\mathrm{CH}}_{3}\mathrm{OH}}$	(cm−2)	$2.4\times {10}^{19}$	≥1 × ${10}^{19}$	$1\times {10}^{19}$	≥1 × ${10}^{18}$
Source size	$(^{\prime\prime} )$	0.22	0.20-0.24	0.24	0.22–0.30
	Predictions versus Millimeter Observations
${T}_{b}{{dV}}_{\mathrm{pred}}$	(K km s−1)	4.7(0.8)		9.1(1.2)
${T}_{b}{{dV}}_{\mathrm{obs}}$	(K km s−1)	$\leqslant$ 0.9		6.5(1.9)
${\tau }_{\mathrm{dust}}^{143\mathrm{GHz}}$		≥1.6		0.3

Note. Bottom: Comparison of the LVG model predictions with the Taquet et al. (2015) millimeter observations (see the text).

Download table as:  ASCIITypeset image

Adopting the 1σ range of gas temperature and density derived for 4A1 and 4A2 (Table 2), we ran a new grid of LVG models with the CH₃OH column density from 1 × 10¹⁸ to 8 × 10¹⁹ cm⁻² at 143–146 GHz to predict the methanol line intensities observed by Taquet et al. (2015). We then used the CH₃OH ${3}_{1}-{2}_{1}$ A⁺ line at 143.866 GHz, which provides the most stringent constraint to the dust optical depth, to compare the predicted intensity with that observed by Taquet et al. (2015). In the comparison, we took into account our LVG-derived source size and the angular resolution of the Taquet et al. (2015) observations. While for 4A2 we considered the line intensity quoted by Taquet et al. (2015), for 4A1, not having CH₃OH detection, we used the 3σ level of the Taquet et al. (2015) observations integrated over 3 km s⁻¹ (average linewidth toward 4A1: see Section 3).

The 4A1 and 4A2 CH₃OH predicted and observed values are reported in Table 2. While the two intensities are similar toward 4A2, they differ by about a factor five toward 4A1.

Assuming that the difference between the predicted and observed intensities is due to the (foreground) dust absorption and using Equation (1), we derived the dust optical depth at 143 GHz (${\tau }_{\mathrm{dust}}^{143\mathrm{GHz}};$ Table 2). While ${\tau }_{\mathrm{dust}}^{143\mathrm{GHz}}$ toward 4A2 is small (∼0.3), that toward 4A1 is large ($\geqslant 1.6$) enough to attenuate the methanol line intensity by a factor ≥5. Therefore, the dust is affecting the millimeter line emission differently in the two sources.

So far, only about a dozen hot corinos have been detected (Section 1) and the question arises whether this is because they are rare or because the searches have always been carried out at millimeter wavelengths, where dust could heavily absorb the line emission.

Our first result is that a source that was not supposed to be a hot corino based on millimeter observations, IRAS 4A1 (López-Sepulcre et al. 2017), indeed possesses a region with temperature $\geqslant 100$ K (Section 3), namely the icy mantle sublimation one, and shows methanol emission (Section 3), the simplest of the iCOMs, when observed at centimeter wavelengths. According to its definition (Ceccarelli 2004), thus, IRAS 4A1 is a hot corino.

Although we cannot affirm that hot corinos are ubiquitous, it is clear that the searches at millimeter wavelengths may be heavily biased and that complementary centimeter observations are necessary to account for dust opacity and understand the occurrence of hot corinos.

Unlike 4A2, no sign of iCOM millimeter emission was revealed toward 4A1 (Taquet et al. 2015; López-Sepulcre et al. 2017). Using ALMA observations at 250 GHz, López-Sepulcre et al. (2017) found that the iCOMs abundances toward 4A2 and 4A1 differ by more than a factor 17, with the largest values (∼100) for HCOOCH₃ and CH₃CN.

The first question to answer is whether the chemical difference between the two coeval objects is real or due to a different absorption by the surrounding dust.

In Section 4.3, we found that ${\tau }_{\mathrm{dust}}$ at 143 GHz toward 4A1 and 4A2 is $\geqslant 1.6$ and 0.3, respectively (see Table 2). Using the dependence of ${\tau }_{\mathrm{dust}}$ from the frequency (${\tau }_{{\nu }_{2}}/{\tau }_{{\nu }_{1}}={\left({\nu }_{2}/{\nu }_{1}\right)}^{\beta }$) and assuming $\beta =2$ (ISM value), the optical depth scaled at 250 GHz (frequency at which López-Sepulcre et al. 2017, derived the above iCOMs abundance ratios) is $\geqslant 4.9$ for 4A1 and 0.9 for 4A2. Therefore, the different dust absorption toward 4A1 and 4A2 provides us, as a lower limit, a factor of 55 difference in their line intensities (${I}^{{\rm{A}}2}/{I}^{{\rm{A}}1}$), comparable to the 4A2/4A1 iCOMs abundance ratios derived by López-Sepulcre et al. (2017). A large dust absorption was also suggested by the anomalous flattened continuum spectral index at 100–230 GHz (Li et al. 2017) and the 90° flipping of the linear polarization position angles observed at above and below 100 GHz frequencies (Ko et al. 2020).

Although we cannot exclude that a real chemical difference exists between 4A1 and 4A2, the observations so far available cannot support that hypothesis. Centimeter observations of other iCOMs than methanol are necessary to settle this issue. This conclusion may apply to other binary systems where an apparent chemical difference is observed using millimeter observations.

The dust absorption also affects the iCOM line intensities in 4A2. At 143 GHz ${\tau }_{\mathrm{dust}}$ is 0.3, which leads to underestimate the iCOM abundances by about 30%. At higher frequencies, this factor becomes more important; e.g., at 250 GHz, where several hot corinos studies are carried out (see references in Section 1), the absorption factor would be 2.5, and at 350 GHz, frequency where the most sensitive iCOMs search has been carried out (e.g., Jørgensen et al. 2016), the absorption factor would be 6. This behavior also agrees with what has already been found in massive hot cores (e.g., Rivilla et al. 2017). Therefore, in order to derive reliable iCOM abundances, complementary centimeter observations are needed to estimate the dust absorption.