Weak and Compact Radio Emission in Early High-Mass Star Forming Regions: II. The Nature of the Radio Sources

TL;DR

This study investigates the nature of weak, compact radio sources in early high-mass star-forming regions, providing evidence that many are ionized jets or pressure-confined HII regions, supporting jet-based models of early stellar ionization.

Contribution

It offers new observational evidence that a significant fraction of early high-mass star radio sources are ionized jets, aligning with recent theoretical models and extending understanding of star formation processes.

Findings

Approximately 30% of sources are ionized jets.

Radio luminosity correlates strongly with bolometric luminosity.

Supports jet-based models of early high-mass star ionization.

Abstract

In this study we analyze 70 radio continuum sources associated with dust clumps and considered to be candidates for the earliest stages of high-mass star formation. The detection of these sources was reported by Rosero et al. (2016), who found most of them to show weak (${\scriptstyle <}$1 mJy) and compact (${\scriptstyle <}\,$0.6$^{\prime \prime}$) radio emission. Herein, we used the observed parameters of these sources to investigate the origin of the radio continuum emission. We found that at least $\sim 30\%$ of these radio detections are most likely ionized jets associated with high-mass protostars, but for the most compact sources we cannot discard the scenario that they represent pressure-confined HII regions. This result is highly relevant for recent theoretical models based on core accretion that predict the first stages of ionization from high-mass stars to be in the form of jets. Additionally, we found that properties such as the radio luminosity as a function of the bolometric luminosity of ionized jets from low and high-mass stars are extremely well-correlated. Our data improve upon previous studies by providing further evidence of a common origin for jets independently of luminosity.

Tables

Table 1. Table 1: Extended Sources: Parameters from Radio Continuum

Region	Radio	$\nu$	Sν	$\Delta s$	EM/${10}^5$	n${}_e{}^{}/{10}^3$	U	log N${}_{Ly}{}^{}{}_{}{}^{\prime }$	Spectral
	Source	(GHz)	($\mu$Jy)	(pc)	(pc cm-6)	(cm-3)	(pc cm-2)	(s-2)	Typeaafootnotemark:
18470$-$0044	C	25.5	3490	0.085	8.5	3.2	9.2	46.4	B0.5
18521+0134	B	25.5	378	0.057	2.5	2.1	4.7	45.5	B1
19035+0641	B	25.5	2270	0.009	39.6	21.1	3.4	45.1	B1
20293+3952	Cbbfootnotemark:	25.5	1560	0.019	1.9	3.2	2.1	44.4	B2
20343+4129	A	25.5	881	0.005	17.5	18.5	1.8	44.3	B2

Table 2. Table 2: Ionized Jets

Region	Radio Source	Jet Direction	Outflow Direction	H${}_2{}^{}-$Jet Direction	New Detection	Reference
G11.11$-$0.12P1	A, C, D	NE$-$SW	E$-$W, NE$-$SWaafootnotemark:	E$-$W	y	(1) (2) (3)
18089$-$1732	A	N$-$S	N$-$S	no/very weakbbfootnotemark:	n	(4) (5) (6)
18151$-$1208	B	NE$-$SW	NW$-$SEccfootnotemark:	NW-SE	n	(7) (8) (9) (10)
18182$-$1433	A$-$Cddfootnotemark:	E$-$W	NE$-$SW, NW$-$SE	E$-$W	n	(11) (12) (13)
IRDC18223$-$3	A$-$Beefootnotemark:	NE$-$SW	NW$-$SEfffootnotemark:	SE-NW	y	(14) (15)
G23.01$-$0.41	A	NE$-$SW	NE$-$SW	non-detection	n	(16) (17) (18) (2)
18440$-$0148	A	NW-SE	$\cdots$ ggfootnotemark:	non-detection	y	(19)
18566$+$0408	A$-$Dhhfootnotemark:	E$-$W	NW$-$SE	non-detection	n	(20)(21) (22) (2)
19035$+$0641	A	NE$-$SW	NW$-$SE	no/very weakbbfootnotemark:	y	(23)
19411$+$2306	A	NE$-$SW	NE$-$SW	detectionbbfootnotemark:	y	(24)
20126$+$4104	A$-$B	NW$-$SE	NW$-$SE, S$-$N	NW$-$SE	n	(25) (26) (27) (28) (29)
20216$+$4107	A	NE$-$SW	NE$-$SW	NE$-$SW	y	(2) (3)

Table 3. Table 3: Ionized Jet/Wind Candidates

Region	Radio Source	Outflow Direction	H${}_2{}^{}-$Jet Direction	Reference
UYSO1	A	NW$-$SE	$\cdots$	(1)
18264$-$1152	F	NW$-$SE	E$-$W	(2) (3)
18345$-$0641	A	NW$-$SE	very weak	(4) (5) (6)
18470$-$0044	B	E$-$W	no/very weakaafootnotemark:	(3)
18517$+$0437	A	N$-$S	very weak	(7) (3)
18521$+$0134	A	$\cdots$ bbfootnotemark:	non-detection	(8)
G35.39$-$00.33mm2	A	$\cdots$	$\cdots$	$\cdots$
18553$+$0414	A	$\cdots$ ccfootnotemark:	non-detection	(3)
19012$+$0536	A	NE$-$SW	non-detection	(3) (3)
G53.25$+$00.04mm2	A	$\cdots$	$\cdots$	$\cdots$
19413$+$2332	A	$\cdots$ ddfootnotemark:	$\cdots$	(3)
20293$+$3952	Eeefootnotemark:	NE$-$SW	detection	(9) (10) (3)
20343$+$4129	B	E$-$W	non-detection	(11) (3)

Equations

$$R_s=4104.7{\left(\frac{N_{Ly}}{10^{49}{s}^{-1}}\right)}^{1/3}{\left(\frac{n_{H_2}}{10^{5}c{m}^{-3}}\right)}^{-2/3}\text{[au]}.$$

$$R=R_s{\left(\frac{2k\xi T_{H^{+}}}{m_{H_2}{\sigma }_v^2+k{T}_k}\right)}^{2/3},$$

$$\left(\frac{{\dot{P}}_{Curiel+}}{M_{\odot }{\text{yr}}^{-1}\text{km}{\text{s}}^{-1}}\right)=\frac{3.13\times 10^{-4}}{\eta }\left(\frac{S_{\nu }{d}^2}{{\text{mJy kpc}}^2}\right){\left(\frac{v_{\star }}{200{\text{km s}}^{-1}}\right)}^{0.32}{\left(\frac{T}{10^4\text{K}}\right)}^{-0.45}{\left(\frac{\nu }{5\text{GHz}}\right)}^{0.1}\left(\frac{\tau }{1-e^{-\tau }}\right),$$

$$\alpha =2+\frac{1}{\text{ln}\left(\frac{\nu }{{\nu }^{\prime }}\right)}\text{ln}\left[\frac{1-\frac{2}{{\tau }^2}\{1-(\tau +1)e^{-\tau }\}}{1-\frac{2}{{\left[\tau {\left(\frac{\nu }{{\nu }^{\prime }}\right)}^{2.1}\right]}^2}\{1-(\tau {\left(\frac{\nu }{{\nu }^{\prime }}\right)}^{2.1}+1)e^{-\tau {\left(\frac{\nu }{{\nu }^{\prime }}\right)}^{2.1}}\}}\right].$$

$$\left(\frac{{\dot{P}}_{Reynolds}}{M_{\odot }{\text{yr}}^{-1}\text{km}{\text{s}}^{-1}}\right)=9.38\times 10^{-6}{\left(\frac{v_{jet}}{100{\text{km s}}^{-1}}\right)}^2\left(\frac{1}{x_0}\right)\left(\frac{\mu }{m_p}\right){\left[\left(\frac{S_{\nu }}{\text{mJy}}\right){\left(\frac{\nu }{10\text{GHz}}\right)}^{-\alpha }\right]}^{0.75}{\left(\frac{d}{\text{kpc}}\right)}^{1.5}{\left(\frac{{\nu }_m}{10\text{GHz}}\right)}^{-0.45+0.75\alpha }$$

$$\times {\theta }_0^{0.75}{\left(\frac{T}{10^4\text{K}}\right)}^{-0.075}(\text{sin}i{)}^{-0.25}{F}^{-0.75}.$$

$$F=\frac{(2.1{)}^2}{q_{\tau }(\alpha -2)(\alpha +0.1)},$$

Full text

Ionized Jets – stars: formation – techniques: high sensitivity – techniques: interferometric

††software: CASA (McMullin et al., 2007), APLpy (Robitaille & Bressert, 2012).

1 Introduction

Young high-mass stars (M $\geq\,$ 8 M⊙) probed in cm-wavelength interferometric studies typically appear as fairly bright (flux densities of $\sim$ few mJy to Jy) regions of ionized gas that are classified according to their size and emission measure, e.g., compact, ultracompact (UC), and hypercompact (HC) HII regions (e.g., Kurtz, 2005). It is generally thought that once nuclear burning has begun the star produces enough UV radiation to photoionize the surrounding gas. However, theories of the earliest stages remain poorly constrained by observations mainly due to the characteristics of the regions where they are born, which are highly dust-obscured, distant ($\gtrsim$ 1 kpc) regions that undergo rapid evolution, and they reach the zero-age main sequence (ZAMS) while still heavily accreting. In fact, an evolutionary sequence for high-mass stars has not yet been established (e.g., Sánchez-Monge et al., 2013a; Tan et al., 2014), although significant progress has been achieved on both observational and theoretical fronts (e.g., Motte et al., 2018). The identification and study of objects in the early stages of their evolution will help us to discriminate among proposed mechanisms for their formation; the two main scenarios being core accretion (i.e., scaled-up version of low-mass star formation) and competitive accretion (i.e., in which stars in a cluster attract each other while they accrete from a shared reservoir of gas; see Tan et al., 2014). The low-mass star formation process is modeled by accretion via a circumstellar disk and a collimated jet/outflow that removes angular momentum and allows accretion to proceed (e.g., Shu et al., 1988). The jet/outflow system is powered magnetohydrodynamically by rotating magnetic fields coupled to either the disk (disk winds: e.g., Konigl & Pudritz, 2000) and/or the protostar (X-winds: e.g., Shu et al., 1987). Additionally, protostellar collisions have been proposed as an alternative mechanism for the formation of high-mass stars (Bonnell et al., 1998; Bally & Zinnecker, 2005).

Massive molecular outflows are a common phenomenon in high-mass star forming regions (e.g., Shepherd & Churchwell, 1996; Beuther et al., 2002b); hence accretion disks and ionized jets similar to those found towards low-mass protostars are also expected. In addition, several surveys toward high-mass star forming regions in the NIR spectral lines of H2 have detected a large number of molecular jets (e.g., Wolf-Chase et al., 2017; Navarete et al., 2015). However, the current sample of known high-mass protostars associated with disks (see review by Beltrán & de Wit, 2016) and collimated jets (e.g., Marti et al., 1995; Martí et al., 1998; Curiel et al., 2006; Rodríguez et al., 2008) is inadequate to draw conclusions about the entire population. The detection of sources at the onset of high-mass star formation and the measurement of their physical properties is essential to test theoretical models of high-mass star formation (e.g., Tan et al., 2014). Furthermore, the most sensitive instruments are necessary to place significant constraints on the occurrence rate and parameters of these detections.

In Rosero et al. (2016, hereafter Paper I) we described our high sensitivity ($\sim$3 – 10 $\mu$Jy beam*-1*) continuum survey, which aimed to identify candidates in early evolutionary phases of high-mass star formation and to study their centimeter continuum emission. We observed 58 high-mass star forming region candidates using the Karl G. Jansky Very Large Array (VLA)111The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. at 1.3 and 6 cm wavelengths at an angular resolution ${\scriptstyle<}\,$0.*′′*6. The 58 targets were grouped into three categories based on their mid and far-IR luminosity as well as the temperature of the cores: 25 hot molecular cores (HMCs), 15 cold molecular cores with mid-IR point source association (CMC–IRs), and 18 cold molecular cores (CMCs) devoid of IR point source associations. The cores in our sample cover a wide range of parameters such as bolometric luminosity and distance. They have similar masses and densities, however, the latter two types of cores—mainly found within infrared dark clouds (IRDCs)—have lower temperatures (T$\sim$ 10–20 K) than HMCs (T${\scriptstyle>}\,$50 K; depending on the probe and scale). In Paper I we reported detection rates of 1/18 (6$\%$) CMCs, 8/15 (53$\%$) CMC-IRs and 25/25 (100$\%$) HMCs. In several cases, we detected multiple sources within a region, which resulted in a total detection of 70 radio sources associated with 1.2 mm dust clumps. The 100$\%$ detection rate of centimeter emission in the HMCs is a higher fraction than previously reported. This suggests that radio continuum may be present, albeit weak, in all HMCs although in many cases it is only detectable with the superior sensitivity now available with the upgraded VLA. Our results show further evidence for an evolutionary sequence in the formation of high-mass stars, from very early stage cold cores (i.e., CMCs) to relatively more evolved ones (i.e., HMCs).

A number of physical processes can cause centimeter continuum emission associated with high-mass star forming regions (see Rodríguez et al. 2012 and Sánchez-Monge et al. 2013b summaries of thermal and non-thermal emission detected at centimeter wavelengths from YSOs). Recently, Tanaka, Tan, & Zhang (2016, hereafter TTZ16) developed a model to predict the radio emission from high-mass stars forming via core accretion. The TTZ16 model predicts that during the first stages of ionization the HII region is initially confined to the vertical (or outflow) axis and produces free-free emission with similar features and parameters as observed towards ionized jets. Ionized jets are detected as weak and compact centimeter continuum sources. At subarcsecond resolutions, they usually show a string-like morphology, often aligned with a large-scale molecular outflow of size up to a few parsecs. Ionized jets trace outflows on smaller scales, providing the location of the driving protostar, that otherwise are deeply embedded in the natal clump and generally remain undetected at other wavelengths due to the high extinction in the region (Anglada et al., 1998). However, less extincted sources may have molecular jet counterparts visible in H2 line emission from shocked gas. These sources are also called ‘thermal radio jets’ due to their characteristic rising spectrum which is consistent with free-free radiation from ionized gas. The ionization mechanism of these jets has been proposed to be shock-induced ionization when the wind from the central protostar ionizes itself through shocks due to variations in velocity of the flow or variations of the mass loss rate (Curiel, Canto, & Rodriguez 1987; Curiel et al. 1989). Unlike the simple model of a uniform electron density HII region, ionized jets and winds have a radial density gradient and thus are partially optically thick. Reynolds (1986) discussed the behavior of collimated jets and the dependency of their physical parameters (such as temperature, velocity, density and ionization fraction) on morphology, independently of the mechanism of ionization, and showed that the spectral index of a partially ionized jet ranges between $-0.1\leq\alpha\leq 1.1$.

The detection of ionized jets toward high-mass stars at their early stages, as predicted by TTZ16, can help to distinguish between accretion scenarios (highly organized outflows are expected from core accretion but not from competitive accretion scenarios; Tan et al. 2016), and ultimately will give us insight about accretion disks around high-mass stars. Several systematic studies searching for ionized jets have been reported in the literature. Guzmán et al. (2012), from a sample of 33 IR luminous objects, detected 2 ionized jets using the Australia Telescope Compact Array (ATCA) with a 4$\sigma$ detection limit and an image rms ($\sigma$) of $\sim$0.1-0.2 mJy beam*-1* at 4.8 and 8.6 GHz. Moscadelli et al. (2016) observed 11 high-mass YSOs using the Jansky VLA and detected 5 collimated ionized jets and 6 ionized wind candidates with a 3$\sigma$ detection limit and an rms $\sim$ 11 $\mu$Jy beam*-1* at $\sim$ 6.2 GHz. Purser et al. (2016) observed 49 high-mass YSOs using the ATCA and detected 16 ionized jets and 12 jet candidates with a 3$\sigma$ detection limit and an rms $\sim$ 17 $\mu$Jy beam*-1* at $\sim$ 5.5 GHz. Additionally, the Protostellar Outflow at the EarliesT Stage (POETS) survey is undertaking a search of radio-jets using the VLA with an angular resolution of $\sim$0.′′1 and an image rms of $\sim$10 $\mu$Jy beam-1 (Sanna et al., 2018, 2019b). In our radio continuum Jansky VLA survey we observed 58 high-mass star forming regions and detected 70 radio sources with a 5$\sigma$ detection limit and an rms $\sim$ 5 $\mu$Jy beam*-1* at $\sim$ 6 GHz (Rosero et al., 2016). Anglada, Rodríguez, & Carrasco-González 2018 is a recent comprehensive review of ionized jets in star forming regions.

The main goal of this paper is to investigate the nature of the 70 detected radio sources reported in Paper I. The observations along with the complete list of targets, coordinates, radio detections and derived observational parameters are presented in Paper I. In Section 2 we examine several scenarios to explain the origin of the ionized gas emission and we study the physical properties of the detected sources. Section 3 contains a discussion of the viability of the different scenarios. In Section 4 we summarize our findings. Additionally, Appendix A shows the bolometric luminosity estimates for these high-mass star forming regions using Herschel/Hi–GAL data and Appendix B shows a study of the momentum rate of ionized jets.

2 Models Considered for the Radio Emission

2.1 Low-mass Young Stellar Objects

The main goal of our high sensitivity continuum survey presented in Paper I was to detect radio emission from high-mass protostars. However, there exists a variety of sources that could also appear as radio detections in our images. In Paper I we considered contamination by extragalactic radio sources, and found that only a small number of extragalactic sources are expected to be observed within the typical dust clump size of $\sim$30*′′* (8 and 2 sources in the 6 and $1.3\,$cm bands, respectively for the entire sample). A more likely source of contamination would be the presence of low-mass YSOs which are expected to be present in regions of high-mass star formation (e.g., Rivilla et al., 2013). We are thus interested in identifying possible low-mass class 0 – class III YSOs that could have been detected in our survey toward high-mass star forming regions.

A large sample of low-mass YSOs has been observed with the VLA at 4.5 and 7.5 GHz as part of the Gould Belt survey (i.e., Ophiuchus at a distance of 120 pc: Dzib et al., 2013; Orion at 414 pc: Kounkel et al., 2014; Serpens at 415 pc: Ortiz-León et al., 2015; Taurus-Auriga at 140 pc: Dzib et al., 2015, and Perseus at 235 pc: Pech et al., 2016). The brightest low-mass YSO in the entire Gould Belt survey (excluding the Orion region) was found in the Ophiuchus region (source J162749.85–242540.5, a class III YSO, i.e., weak-lined T-Tauri star, with S${}_{7.5\,GHz}=8.51\,$mJy, Dzib et al. 2013). To determine whether such an object would have been detected in our survey, we scaled its flux density to the assumed distance222Distances were taken from the literature, and are listed in Table A. Most distances are kinematic; only a few regions have trigonometric parallax measurements. of each of our targets, and compared its scaled flux density to our adopted detection limit of $\geq$ 5 times the image rms at 7.4 GHz for each of our regions. We found that such a YSO would not be detected in any of our targets located at distances beyond $2\,$kpc. Since the majority of our targets exceed this distance (see Figure 1), we conclude that for most of our observed regions the detected radio sources are not low-mass YSOs.

There are 10 regions in our survey that are located at distances $\leq$ 2 kpc. However, given the 7.4 GHz image rms for these regions, only in 7 of them would we have detected the brightest low-mass YSO of Ophiuchus. These regions are five HMCs: 18517$+$0437, 20126$+$4104, 20293$+$3952, 20343$+$4129, G34.43$+$00.24mm1, and two CMC–IRs: LDN1657A$-$3 and UYSO1. In these 7 regions we detected a total of 13 radio sources within the FWHM of the mm clumps: 10 towards HMCs and 3 towards CMC–IRs. That some of these sources are possibly low-mass YSOs can be seen in the case of IRAS 20126$+$4104: Besides the well-studied high-mass protostars associated with radio sources 20126$+$4104 A and 20126$+$4104 B, the radio source G$78.121+3.632$ in this region (see Paper I , Table 4) corresponds to the source I20var, which was discussed by Hofner et al. (2007). This is a highly variable radio source and has observational properties consistent with a flaring T-Tauri star. In the same region, we have also detected a new object of similar characteristics. Radio source 20126$+$4104 C was detected for the first time in our survey although several high sensitivity observations of this region have been made in the past (Hofner et al., 2007).

Hence, 20126$+$4104 C is clearly variable in the radio regime, and is a candidate for a low-mass pre-main sequence star. Additionally, the radio source LDN1657A$-$3 A, which has a negative spectral index ($\alpha=-1.2$), is also a candidate for a variable radio source, where the emission is probably caused by non-thermal processes on the surface of a T-Tauri star. While the observational properties of these sources are consistent with low mass YSOs, we note that alternative explanations are possible (e.g., Cesaroni et al., 2018).

In summary, while some degree of contamination by low-mass YSOs probably exists in our survey for the nearest sources, for the majority of our targets the detected radio sources are very likely not contaminated by emission from low-mass YSOs.

2.2 HII Regions

In Paper I we reported the detection of 70 radio continuum sources associated with three different types of mm clumps and we calculated their 5–25 GHz spectral index ($\alpha$) using power-law fits of the form $S_{\nu}$ $\propto$ $\nu^{\alpha}$. The spectral index values and the fits to the data for all the radio detections are reported in Paper I in Table 4 (electronic version) and in Figure 4, respectively. The range of spectral indices found was $-$1.2 to 1.8 (see Figure 5 in Paper I). Based on their radio spectra, we classify these sources as flat spectral index ($-$0.25$<\alpha<$0.2), positive spectral index ($\alpha\geq$ 0.2), and negative spectral index ($\alpha<$-0.25). Thus, we have 10 sources with flat, 44 sources with positive, and 9 sources with negative spectral index. For the remaining 7, there is not a clear estimate of the spectral index.

The radio sources have a variety of morphologies. Excluding the sources without spectral index information, there are 6 extended sources, 8 sources with elongated structures, and the majority of sources (49) are compact with respect to our synthesized beam. In this section we consider whether a family of HII region models could explain sources with flat and positive spectral index.

2.2.1 Extended Sources

Among the sources detected in our survey associated with $1.2\,$mm dust emission, there are six sources that are clearly extended at cm wavelengths with respect to the $\sim$0.′′4 resolution of the maps, hence they are candidates for HII regions, i.e. photoionized gas. These sources are relatively bright (S${}_{25.5\,GHz}\approx$ 1 mJy), and are found mostly toward HMCs. Moreover, they generally show a flat spectral index, indicative of optically thin free-free emission. For five of these sources, we calculate the physical properties from the $25.5\,$GHz continuum flux using the formulae from Kurtz et al. (1994), which assume spherical symmetry, and optically thin emission from a uniform density plasma with T${}_{e}=10^{4}$ K. The results are listed in Table 2.2.1, where column 1 is the region name, column 2 is the specific radio source, and columns 3 and 4 are the frequency ($\nu$) and radio flux (Sν), respectively. Column 5 is the observed linear size (diameter) of the radio source ($\Delta s$) at 3$\sigma$ rms level in the image, column 6 is the emission measure (EM), column 7 is the electron density (ne), column 8 is the excitation parameter (U) and column 9 is the logarithm of the Lyman continuum flux (N${}^{\prime}_{Ly}$) required for ionization. We use log N${}^{\prime}_{Ly}$ to estimate the spectral type of the ionizing star (listed in column 10) using the tabulation in Panagia (1973), further assuming that a single ZAMS star is photoionizing the nebula and producing the Lyman continuum flux. The distances used for these calculations are listed in Table A, and the near kinematic distance is adopted when the region has a distance ambiguity.