EDP Sciences
Free Access
Issue
A&A
Volume 541, May 2012
Article Number A72
Number of page(s) 9
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201218999
Published online 01 May 2012

© ESO, 2012

1. Introduction

High-angular resolution observations of high-mass star forming sites in the 6.7 GHz methanol maser emission have revealed a diversity of morphological structures from very simple composed of a few milliarcsecond (mas) wide spots to complex and extended clouds of arcsecond sizes (Phillips et al. 1998; Walsh et al. 1998; Minier et al. 2000; Dodson et al. 2004; Bartkiewicz et al. 2009; Pandian et al. 2011). Linear structures with a velocity gradient are thought to be the signature of an edge-on disc or torus around young massive protostars (Norris et al. 1998; Minier et al. 2000) but there is evidence that some of them are associated with outflows (De Buizer 2003) or even a shock propagating through a rotating dense molecular clump (Dodson et al. 2004). Arc-like or ring-like morphologies of methanol emission seem to prove a disc scenario (Bartkiewicz et al. 2009), although most of those structures do not show signs of rotation but rather inflow or/and outflow dominate (van Langevelde et al. 2010), it is then postulated that methanol masers arise in the shock interface between the large scale accretion and a circumstellar disc.

While the 6.7 GHz methanol maser line is radiatively pumped in warm (T ~ 150 K) and dense (n ≤ 108 cm-3) regions (Cragg et al. 2005; Sobolev et al. 1997), the 22 GHz water maser emission is collisionally pumped and probes denser (n ≥ 108 cm-3) and hotter (T ~ 400 K) gas (Elitzur et al. 1989). The water maser is often excited in strong shocks, driven by young low-mass and high-mass (proto)stellar objects, on a working surface between a outflow and a dense envelope. A number of morphologies of this emission is observed from collimated jet, wide-angle flows, expanding shells to equatorial flows (Goddi et al. 2007; Moscadelli et al. 2007, 2011; Sanna et al. 2010a,b; Torrelles et al. 2011). One or more types of those maser structures occur simultaneously in star forming regions on scale sizes of a few mas to several arcsec and is possibly related to the geometry of the envelopes or the mechanisms driving the outflows.

Our recent VLA survey of 22 GHz water masers in a sample of 31 methanol sources has yielded 22 detections (Bartkiewicz et al. 2011) where both maser species are excited by the same central objects. We noted that a distinct group of methanol sources with ring-like structure show either no associated water masers at all or water masers that are distributed orthogonally to the major axis of the ring. It is argued that the methanol maser structure traces a circumstellar disc/torus around a high-mass young stellar object while the water masers originate in outflows. As the VLA data are of moderate angular and spectral resolutions of  ~1′′  and 0.65 km s-1, respectively, we undertook the VLBI observations to examine a disc-outflow scenario in the two brightest water maser targets, which show the arc-like or ring-like morphologies that are characteristic of methanol masers.

The source G31.581 + 00.0771 has been recognized as a massive young stellar object based on 6.7 GHz methanol maser observations (Szymczak et al. 2000). The methanol maser spots are distributed along an arc or ring of 217 mas size (Bartkiewicz et al. 2009), which corresponds to  ~1200 AU for the assumed near kinematic distance of 5.5 kpc (Reid et al. 2009). The detection of an infrared source of bolometric luminosity of 3 × 104   L (Urquhart et al. 2011), 22 GHz H2O maser emission (Bartkiewicz et al. 2011), 1665 and 1667 MHz weak OH masers (Szymczak & Gérard 2004), millimeter molecular thermal lines (Szymczak et al. 2007; Urquhart et al. 2008), and 5 and 8.4 GHz continuum emission at location of 31.582 + 00.075 (i.e., 9′′ apart from the maser arc/ring) (Urquhart et al. 2009; Bartkiewicz et al. 2009) indicates that this is a cluster of recent star formation.

thumbnail Fig. 1

G31.581 + 00.077. a) Spitzer GLIMPSE 4.5–3.6 μm excess image overlaid with the water maser positions (yellow asterisks) from the EVN observation and H II region (yellow cross) detected by Bartkiewicz et al. (2009). b) Shows the distribution of water masers observed with the EVN (filled squares) from this paper and the VLA (open squares) from Bartkiewicz et al. (2011). The colours of the symbols relate to the LSR velocities as indicated on the right-hand side of the plot. The origin of the map is the position of the brightest 6.7 GHz methanol maser spot (Bartkiewicz et al. 2009) (see also Table 1). c) Shows an enlargement of the north-western (NW) water masers. d) Shows an enlargement of the south-eastern (SE) water masers together with the distribution of the 6.7 GHz methanol masers, marked by open circles, from Bartkiewicz et al. (2009). In c) and d), the arrows represent the velocity gradients (from blue- to red-shifted LSR velocities) detected within individual cloud. The black numbers and letters correspond to the clouds with internal velocity gradients, while the grey ones correspond to the clouds without internal velocity gradients (Table 3).

Open with DEXTER

thumbnail Fig. 2

Individual component spectra with Gaussian velocity profiles of 22.2 GHz water masers (red squares and lines) and 6.7 GHz methanol masers (blue circles and lines) towards G31.581 + 00.077 detected using the EVN in 2010 and 2007, respectively. The numbers and letters correspond to the cloud labels as given in Table 1 and Fig. 1. The grey lines show individual Gaussian profiles fitted to the blended features. The clouds with only two single spots are not marked to improve the clarity of the figure (i.e., clouds b, e, g).

Open with DEXTER

The source G33.641 − 00.228 detected in the 6.7 GHz methanol line (Szymczak et al. 2000) has an arc spot distribution of length 630 AU (Bartkiewicz et al. 2009) for the assumed near kinematic distance of 3.8 kpc (Reid et al. 2009), which seems to be more likely than the far kinematic distance (Zhang, priv. comm.). The site also contains water and OH masers (Bartkiewicz et al. 2011; Szymczak & Gérard 2004) but no 8.4 GHz continuum emission was detected with an upper limit of 0.15 Jy (Bartkiewicz et al. 2009). The detection of radio recombination lines at about 103 km s-1 (Anderson et al. 2011) and a multi-feature 13CO line spectrum (Urquhart et al. 2008) illustrates the complexity of this molecular cloud and its clustered star formation.

2. Observations and data reduction

The EVN2 observations of G31.581 + 00.077 and G33.641 − 00.228 with the antennas at Jodrell Bank, Effelsberg, Medicina, Metsähovi, Onsala, and Yebes, were carried out at 22.23508 GHz on 2010 October 30 for 8 h (the project EB047). The tracking phase centres were estimated from the VLA survey for the water maser spots that were located nearest to the methanol emission in each target (Bartkiewicz et al. 2011) at α = 18h48m41951, and α = 18h53m32563, δ =  + 00°31′39130 (J2000), respectively. A phase-referencing scheme was applied with a reference source J1851 + 0035 (from the VLA calibrator catalogue), using a cycle time between the maser and phase-calibrator of 60 s + 90 s. This strategy yielded 2 h on-source times for each target. The bandwidth was set to 8 MHz and data were correlated with the Mk IV Data Processor operated by JIVE with 1024 spectral channels. The resulting spectral resolution was 0.1 km s-1. The velocity was measured with respect to the local standard of rest (LSR).

The data calibration and reduction were carried out with NRAO’s Astronomical Image Processing System (AIPS) using standard procedures for spectral line observations. We used the Effelsberg antenna as a reference. The amplitude was calibrated by performing measurements of the system temperature at each telescope and applying the antenna gain curves. The parallactic angle corrections were subsequently added to the data. The source 3C 454.3 was used as a delay, rate, and bandpass calibrator. The phase-calibrator was imaged and a flux density of 190 mJy was obtained. The maser data were corrected for all Doppler effects and self-calibrated using the brightest and most compact maser spot. Finally, naturally weighted maps of spectral channels were created in the velocity range where the emission was seen in the scalar-averaged spectrum. For imaging, the resulting synthesized beam was 1.0    ×    2.4 mas2 with a position angle (PA) of  − 38°, while the pixel separation was 0.2 mas in both coordinates. The rms noise levels (1σrms) in line-free channels was typically 10 mJy beam-1. The weakest detected maser spot had a brightness of 94 mJy beam-1 that is more than 9σrms.

The positions of water maser spots in all channel maps were determined by fitting two-dimensional Gaussian models. The formal fitting errors resulting from the beamsize/signal-to-noise ratio were smaller than 0.1 mas. To determine the position accuracy of registered maser spots, we need to consider the following factors (Diamond et al. 2003): i) the uncertainty in the phase-reference source position of 2 mas3; ii) the antenna positions, where the claimed accuracy of  ~ 1 cm corresponds to 1 mas in RA and 2 mas in Dec; iii) the phase transfer over the separation between targets and the phase-calibrator: 192 (between G31.581 + 00.077 and J1851 + 0035) and 045 (between G33.641 − 00.228 and J1851 + 0035). These caused potential phase-solution transfer errors corresponding to 1 mas in RA and 2 mas in Dec for G31.581 + 00.077 and to 0.2 mas in RA and 0.4 mas in Dec for G33.641 − 00.228, respectively. In total, the absolute position accuracy (1σpos) is 2.5 mas in RA and 3.5 mas in Dec for both targets.

3. Results

Water maser emission towards both targets was mapped, after successful phase-referencing, in the areas of 5′′ ×  5′′  and the velocity range explored by Bartkiewicz et al. (2011) using the VLA. The EVN maps of water maser spots (Figs. 1 and 3) are overlayed on the VLA maps and the 6.7 GHz methanol maser distributions obtained with the EVN (Bartkiewicz et al. 2009). The Spitzer GLIMPSE maps4 of the 4.5–3.6 μm emission excess are added. Following the procedure described in Bartkiewicz et al. (2009), the water maser spots are analysed to identify individual velocity-coherent maser clouds. Their basic parameters, such as positions, ΔRA, ΔDec, LSR velocities, VLSR, and intensities of the brightest spot, Sp, of each cloud are listed in Tables 1 and 2. The water and methanol maser spectra of single clouds are combined in Figs. 2 and 4 with overlays of individual Gaussians, if emission was seen in at least three consecutive channels. A Gaussian analysis of individual clouds is carried out and the fitted flux amplitude, Sfit, the full width at half-maximum, FWHM, the projected extent between the most separated single spot centres within a cloud, Lproj, and the velocity gradient, Vgrad are listed in Table 3. The lower limit to the brightness temperature, Tb, of each cloud is also calculated according to Eq. (9−27) of Wrobel & Walker (1999). For comparison purposes, the same analysis is done for the 6.7 GHz methanol maser data (Table 3) obtained in 2007 June (G31.581+00.077) or 2003 June (G33.641−00.228) with the EVN. Below we comment in more detail on each source.

3.1. G31.581+00.077

A total of 91 water maser spots were detected. They are concentrated in two distinct regions (Fig. 1). The south-east (SE) region of size 50 × 30 mas2 containing weak (<1.35 Jy) emission in the velocity ranges from 90.1 to 93.9 km s-1 and from 100.0 to 103.3 km s-1 is located close to the phase centre. The second region (NW) of middle velocity emission (96.8−100.7 km s-1) and flux density of 2.6−16.5 Jy is located offset by  ~5′′  to the north-west (PA  =  −60°) of the phase centre.

The SE region is divided into the cluster of clouds 1, 2, and 3 separated by 50 mas from the cluster of clouds 6, 7, and 8 (Fig. 1). The clouds in the SE region show Gaussian velocity profiles with FWHM linewidths of 0.54 − 0.94 km s-1. Their projected sizes are 0.07 − 0.47 mas, which correspond to 0.9−2.6 AU for the adopted distance of 5.5 kpc. The brightness temperature is 0.24 − 1.38 × 109 K. We note that clouds 3 and 6 show the velocity gradient of 1.3−1.8 km s-1 mas-1 oriented along PAs of 1° and  − 23°, respectively. It is remarkable that the two clusters of clouds (1, 2, 3 and 6, 7, 8) appear very close in space (Table 1) forming arc-like filaments of projected length of 4 and 3.2 mas corresponding to 22 and 18 AU, respectively, and elongated at PA =  −58°. The closest clouds 3 and 6 are separated by 285 AU and a velocity difference of 6.7 km s-1. The mean linear separation between the two clusters along PA = 65°  is 310  ±  8 AU, while the velocity spread is 13.2 km s-1. The velocities of the two filaments are approximately symmetric with the regard of the systemic velocity of 96.0 km s-1 (Szymczak et al. 2007). The two SE filaments are likely signatures of flattened shock surfaces (Torrelles et al. 2001). We found that the water cloud 7 coincides within  ± 0.1 km s-1 with the methanol maser cloud k, while the spatial separation is 24.4 mas which corresponds to the projected distance of 134 AU. Although these two lines were observed within a time span of three years their separation seems to be real. This confirms that the water and methanol masers probe different parts of the environment of young massive stars (e.g., Beuther et al. 2002; Sanna et al. 2010a,b) because of the different pumping mechanisms affecting both species (Elitzur et al. 1989; Cragg et al. 2005).

The NW region of elongation of 26 mas is composed of two clouds 4 and 5 of brightness temperatures of 2.7−17.1 × 109 K (Table 3). Their velocity profiles are obviously asymmetric and nicely fitted by the sums of three and two Gaussian components (Fig. 2) of FWHM linewidth of 0.5 − 0.9 km s-1. The velocities of the NW water maser emission largely coincide with those of the methanol masers but their linear separation is about 27 000 AU. The map of 4.5–3.6 μm emission excess (Fig. 1), which possibly traces shocked molecular gas in outflows from massive stars (e.g., Davies et al. 2007), shows that the SE and NW maser clusters lie in a very complex region. Both maser clusters are likely associated with different powering sources.

Table 1

List of 22.2 GHz water maser clouds as found in EVN observations towards G31.581 + 00.077.

Table 2

List of 22.2 GHz water maser clouds as observed using EVN towards G33.641 − 00.228.

The comparison of our maps with those obtained about 14 months earlier with the VLA in a CnB configuration (Bartkiewicz et al. 2011) implies that the cloud 1 coincides to within 13 mas, the cloud 4 to within 56 mas, and the cloud 7 to within 64 mas with the matching VLA spots. The emission seen with the VLA from three components located  ~ 1 − 2′′  eastward of the clouds 4 and 5 is not detected. We also note that the water maser components detected have very similar amplitudes in both EVN and VLA observations.

3.2. G33.641−00.228

Twenty-one water maser spots were detected in the velocity range of 54.4−57.6 km s-1 over the north-south elongated area 50 × 280 mas2 close to the phase centre (Fig. 3). Using the above-mentioned procedure, we found four maser clouds where the emission is seen in at least three contiguous spectral channels. There are also two other clouds (5 and 6) that do not obey this criterion but seem to be real (Table 2). All but one of the cloud are in the southward cluster at a distance  ~270 mas from the phase centre. This filament cluster of size 41.6 mas is aligned along a PA of 78°. For the adopted distance of 3.8 kpc, the corresponding linear scales are 1020 AU and 157 AU. The clouds 14 in the southern cluster have Gaussian profiles with FWHM linewidths of  0.25 − 0.88 km s-1. We note that none of the water clouds have an internal velocity gradient. The linear size of the individual clouds is 0.3 − 1.0 AU and the brightness temperature is 0.12−0.27 × 109 K (Table 3).

The strongest emission (Tb = 0.63 × 109K) comes from cloud 2, which is a blend of two Gaussian components (Fig. 4). Cloud 2 is  ~20 mas (75 AU) away from the brightest methanol maser clouds a and r and differ in terms of velocity by 1.5 km s-1. This is probably the first such tight association of methanol and water masers. We note that the emission of cloud 2 is blue-shifted by 4.2 km s-1 and that of cloud a is red-shifted by 1.2 km s-1 relative to the systemic velocity of 61.5 km s-1 (Szymczak et al. 2007). It is therefore possible that these clouds signpost a shock front in which the water emission originates behind the front, while the methanol emission (cloud a) appears outside of the shock interface in the infalling gas.

The image of the 4.5 μm−3.6 μm emission excess (Fig. 3) shows that the water maser emission is located 12 to the south-east of single mid-infrared object. This is likely the source powering the outflow along a PA of 165°  traced by the water masers.

The positions of the clouds 2, 3, 4 and 6 differ by about 92 mas from the VLA positions of the matching clouds (Bartkiewicz et al. 2011), which is well within the  ~150 mas accuracy of the CnB configuration VLA data. A weak (<0.9 Jy) red-shifted emission in the velocity range of 83.8 − 85.1 km s-1 seen with the VLA towards cloud 2 (Fig. 1) was not recovered in the VLBI observation. The brightness of the emission from cloud 2 determined in the EVN observation is only slightly lower than that measured with the VLA. In contrast, the maser spots from the southern region (clouds 1, 3 − 6) are about ten times weaker in the EVN maps that may suggest the presence of diffuse and/or highly variable emission.

Table 3

Parameters of 22.2 GHz water and 6.7 GHz methanol maser clouds.

thumbnail Fig. 3

Same as Fig. 1 but for G33.641 − 00.228.

Open with DEXTER

thumbnail Fig. 4

Same as Fig. 2 but for G33.641−00.228. The scale on the left side describes the water maser intensity, while the right side corresponds to the methanol maser intensity. The clouds with only two single spots or without Gaussian characteristic are not marked for the clarity of the figure (i.e., clouds 5, 6, e, g, i, n, p).

Open with DEXTER

4. Discussion

4.1. Maser environments

The Spitzer GLIMPSE map (Fig. 1) clearly indicates that the source G31.581+00.077 lies in a large (35′′ × 25′′) complex area with an 4.5 μm emission excess in the form of clumped and diffuse structures. This extended emission is proposed to be a tracer of shocked gas from regions where outflowing gas interacts with the surrounding medium (Davis et al. 2007; Cyganowski et al. 2009). The 13CO(1−0) spectrum obtained with a 46′′ beam (Urquhart et al. 2008) towards the position 31.5808+0.0757 has two Gaussian components centred at 96.2 and 109.8 km s-1. A weak emission line is clearly detected in the range of 85 to 105 km s-1, i.e., in the wings of the first component suggests the presence of outflows. A signature of ordered motions also appears in the HCO+(1−0) spectrum (Szymczak et al. 2007), where in this slightly asymmetric profile the strongest emission is blueward of the source systemic velocity and can be interpreted as the result of infall (e.g., Fuller et al. 2005). We then find 92 east of the SE water maser an H II region of flux density 10.3 and 15 mJy at 5 GHz (White et al. 2005) and 8.4 GHz (Bartkiewicz et al. 2009), respectively. This continuum source coincides to within 06 with a clump of 3.6–4.5 μm emission excess (Fig. 1). The NW water masers are located at the borders of two clumps of excited gas, and their powering source is unclear although it seems unlikely that the SE and NW masers are associated with the same central object. A comparison of the present EVN maps with those obtained with the VLA 14 months earlier (Bartkiewicz et al. 2011) implies that there is a lack of water masers eastward of the NW clouds (Fig. 1). It is likely that these are diffuse and low intensity masers resolved out with the EVN beam. The spectrum obtained with a 40′′  beam sometime between November 2009 and December 2010 (Urquhart et al. 2011) indicates that the peak flux density of 125 Jy at 99.7 km s-1 is about one order of magnitude higher than that measured with the EVN. We suggest that the water masers of G31.581+00.077 are produced by different young stellar objects in a complex region composed of at least a few high-mass stars well-signposted by the radio continuum and the methanol and water masers.

The water maser in G33.641 − 00.228 lies only 12 south-eastward of the maximum of the 3.6–4.5 μm emission excess of cometary-like morphology (Fig. 3). The HCO+(1−0) spectrum obtained towards the position 33.648 − 0.224 (Szymczak et al. 2007) shows a small dip near 61.5 km s-1 and a slight redward asymmetry that is indicative of outflow motions. The less optically thin H13CO+(1−0) transition is detected as marginal red-shifted emission, which is consistent with an outflow. In the region where the OH 1665, 1667, and 1720 MHz masers were detected (Szymczak & Gérard 2004) and all of them peak at 60.2 km s-1. The strongest 1720 MHz emission and a broad (7.4 km s-1) 1667 MHz absorption profile near 56.2 km s-1 are indicative of shock fronts and a continuum background source, respectively. No continuum emission at 8.4 GHz was detected with a 3σrms limit of 0.15 mJy beam-1 (Bartkiewicz et al. 2009). A weak (~20 mJy) 8.6 GHz hydrogen radio recombination-line near 102.9 km s-1 detected at position G033.645 − 0.227 (Anderson et al. 2011) differs so greatly in terms of velocity from the water masers that this may be a chance projection that is unassociated with the methanol and water masers. We argue that the water masers of G33.641 − 00.228 as well as the associated methanol and hydroxyl masers are excited by individual high-mass stars.

4.2. Physical parameters of circumstellar medium

Lower limits to the brightness temperature of the SE water maser components associated with G31.581 + 00.077 are always lower than 1.4 × 109 K, while the measured linewidths range from 0.54 to 0.94 km s-1 are quite commonly in the star-forming sources (e.g., Goldreich & Kwan 1974; Surcis et al. 2011a; 2011b). Since the kinetic temperature of the masing gas might be expected to be about 400 K (Elitzur et al. 1989), the intrinsic thermal linewidth given by , where Tk is a kinetic temperature, k the Boltzmann constant, and m is a molecular mass should be  ~ 1 km s-1. This value is larger than the observed linewidths and suggests that the masers are unsaturated. In the same volume of gas where the water maser cloud 7 and the methanol maser cloud k appear to coincide (Fig. 1), the methanol thermal linewidth would be 0.75 km s-1, whereas the measured value is 0.30 km s-1. The narrowing of the line profile is expected when the maser is unsaturated. Detailed calculations have shown that in one of the strongest known methanol sources NGC 7538 about 92% of the components are unsaturated (Surcis et al. 2011b). In the source G33.641 − 00.228, the observed linewidths of methanol components are generally narrower than in G31.581 + 00.077, which suggests that the maser is also unsaturated in this region.

The mas-scale velocity gradients in both lines are observed in G31.581 + 00.077. We note that the velocity gradient at 22 GHz is about one order of magnitude higher than at 6.7 GHz (Table 3). Extensive discussion of the velocity gradients for methanol masers by Moscadelli et al. (2011b) suggests that there is a kinematical interpretation of their origin. We note that the methanol maser gradients in both targets can reflect the ordered motions on scales of 10 − 60 AU. Large velocity gradients for water masers in G31.581 + 00.077 suggest that they are related to the outflow motions with velocities larger than 50 km s-1 (see Sect. 4.3).

Table 4

Parameters derived by fitting the kinematics of the rotating and expanding disc model.

4.3. Kinematic models

The two sources investigated in this study possibly belong to a group of objects where a ring-like or arc-like methanol maser distribution traces a circumstellar disc/torus around a high-mass young stellar object, whereas the water maser distribution is orthogonal to the major axis of the methanol structure (Bartkiewicz et al. 2011). Furthermore, these objects are not associated with detectable continuum emission at cm wavelengths (Bartkiewicz et al. 2009) and may represent an early stage of evolution. The new EVN observations of both sources have shown that all of the individual water maser spots detected previously with the VLA (Bartkiewicz et al. 2011) unfold into complex and filament structures of sizes 18 − 160 AU. To examine a disc-outflow scenario in the two sources, we used a model of a rotating and expanding thin disc (Uscanga et al. 2008) for the methanol masers and a model of the outflow (Moscadelli et al. 2000) for the water masers. Detailed descriptions of the modelling are given in the above cited works, and elsewhere for methanol masers in discs (Bartkiewicz et al. 2009) and outflows (Bartkiewicz et al. 2006). We note that in the following we analyse the new water maser data with the methanol maser data obtained 3 − 7 years earlier (Bartkiewicz et al. 2009). However, our inspection of a few sources published so far and systematic EVN monitoring demonstrate that the overall methanol maser morphologies are stable on a scale time of 6 − 8 years (Bartkiewicz et al., in prep.).

thumbnail Fig. 5

Velocity of the methanol maser spots (open circles) in G31.581 + 00.077 and G33.641 + 00.228 versus azimuth angle measured from the major axis. The sinusoidal line represents the best-fit kinematical model of a rotating and expanding disc of infinitesimal thickness to the methanol spots with the parameters listed in Table 4. For completeness, the water maser spots are shown as squares. In a case of G33.641 + 00.228 two kinematical models are presented for a powering source located as follow: D1 – in the centre of the best-fitted ellipse, D2 – at the MIR position.

Open with DEXTER

In Tables 4 and 5, we summarize the best-fit values of both models. For the rotating and expanding thin disc, the rotation (Vrot), expansion (Vexp), and systemic (Vsys) velocities as well as the inclination angle, i, i.e., the angle between the line-of-sight and the normal to the ring plane defined to be is given for each source. The solutions were based on the minimization of the function given by Eq. (8) in Uscanga et al. (2008). The outflow model for water masers is characterized by the vertex of the cone (Xo, Yo), the x-axis coinciding with the projection of the outflow on the plane of the sky at the position angle PA, the inclination angle between the outflow axis and the line-of-sight (i.e., the z-axis) Ψ, and the opening angle of the outflow/cone 2θ. The systemic LSR velocities, Vc, were assumed to be the same as given in Sect. 3. The χ2 function was assumed to be expressed as in Eq. (3) of Moscadelli et al. (2000).

The methanol emission in G31.581 + 00.077 is most accurately reproduced by the model where the powering source is at the centre of the best-fit ellipse and the rotation velocity of 0.9 km s-1 is a factor of two lower than the expansion velocity (Table 4, Fig. 5). These values are typical of the class of ring-like methanol masers (Bartkiewicz et al. 2009) and suggest that the methanol structure of  ~1000 AU diameter cannot be interpreted as a Keplerian disc. It is instead proposed that the methanol masers arise on the interface region between a large-scale accretion flow and a stellar disc (van Langevelde et al. 2010; Torstensson et al. 2011). We also note that the best-fit systemic velocity of 98.54 km s-1 corresponds well to that estimated from molecular lines (Sect. 3.1). The SE water emission is reasonably well-fitted by two different models. In the first model, O1, the vertex of the cone with 2θ = 30°  coincides with the centre of the methanol ellipse within 25 mas and the velocity of masers is 77 km s-1 (Table 5, Fig. 6). We argue that this fit is consistent with the outflow scenario.

As we failed to identify any mid-infrared (MIR) counterpart to the centre of the methanol ellipse as a powering source of the methanol structure and perpendicular to the water maser outflow, we propose an alternative model. This second model, O2, assumes that the position of the cone vertex is just between the red-shifted and blue-shifted voids of SE water masers. The best-fit velocity of the outflow is  ~8 km s-1, while the projection of the axis outflow onto the plane of the sky is roughly parallel to the main axis of the methanol structure (Table 5, Fig. 6). In star forming regions, different centres of activity are reported on scale sizes of a few to several hundred AUs (e.g., Torrelles et al. 2001; 2011). The proper motion studies of G31.581 + 00.077 can clearly provide conclusive evidence for or against this hypothesis.

In the case of methanol masers in G33.641 − 00.228, we present the two kinematical models for a powering source located at the centre of the best-fit ellipse (ΔRA, ΔDec) = (59 mas,17 mas) and that located at the MIR position (−495 mas, 380 mas). In the first model, D1, the infall velocity of about 1 km s-1 is slightly higher than the rotation velocity (Table 4). The structure of water masers is fitted by a narrow ouflow of 2θ = 18°  and velocity of 15 km s-1 (Table 5, Fig. 6). It is consistent with a jet-like outflow of water masers roughly perpendicular to the main axis of methanol structure. In the second model, D2, the visible methanol structure is only a small part of the disc radius of  ~1900 AU and the rotation velocity of 3.7 km s-1 (Table 4), which implies that the enclosed mass is about 30 M. In this case, the water masers may trace the wind gas from a disc.

We note that the kinematic models are unable to accurately reproduce the observed structures of the methanol and water masers in both targets, and that only measurements of the proper motions of maser spots of the two species can unambiguously constrain the proposed scenarios.

Table 5

Parameters derived from fitting the biconical outflow model of Moscadelli et al. (2000).

thumbnail Fig. 6

Outflow models fitted to water (squares) in G31.581 + 00.077 and G33.641 + 00.228 according to the model of Moscadelli et al. (2000). The relevant parameters are listed in Table 5. The right panel presents a comparison of obtained data (squares) vs. model (crosses). Vc is the systemic LSR velocity as given in Sect. 3.

Open with DEXTER

5. Conclusions

We have carried out high-angular resolution studies of the 22.2 GHz of water maser line towards two methanol maser sources G31.581 + 00.077 and G33.641 − 00.228 using the EVN. Although their morphologies did not differ significantly from the previous VLA results, the astrometry at the mas level and the properties of the maser clusters could be estimated owing to the high-angular and spectral resolution. In total, we detected eight and six water maser clouds towards G31.581 + 00.077 and G33.641 − 00.228, respectively. In the first target, the water maser components are associated with different centres of star-forming activity, and the components associated with the methanol ring-like structure possibly trace the outflow. In the source G33.641 − 00.228, southern water masers possibly trace a wind from a disc. The kinematic models containing ring-like or arc-like methanol maser structures are able to trace a circumstellar disc/torus around a high-mass young stellar object, whereas the water maser distribution is orthogonal to the major axis of the applied methanol structure and poorly constrained by the present data. The present studies show that the two sources are good targets for proper motion studies in order to understand more clearly the kinematics of gas in the environments of high-mass stellar objects. They also encourage us to extend the multi-epoch EVN observations for a whole sample.


1

The names of two targets follow the Galactic coordinates of the brightest methanol maser spots derived by Bartkiewicz et al. (2009).

2

The European VLBI Network is a joint facility of European, Chinese, South African and other radio astronomy institutes funded by their national research councils.

3

The GSFC ICRF2 VLBI Source Position Catalog.

Acknowledgments

A.B. and M.S. acknowledge support by the Polish Ministry of Science and Higher Education through grant N N203 386937. This work has also been supported by the European Community Framework Programme 7, Advanced Radio Astronomy in Europe, grant agreement No.: 227290.

References

All Tables

Table 1

List of 22.2 GHz water maser clouds as found in EVN observations towards G31.581 + 00.077.

Table 2

List of 22.2 GHz water maser clouds as observed using EVN towards G33.641 − 00.228.

Table 3

Parameters of 22.2 GHz water and 6.7 GHz methanol maser clouds.

Table 4

Parameters derived by fitting the kinematics of the rotating and expanding disc model.

Table 5

Parameters derived from fitting the biconical outflow model of Moscadelli et al. (2000).

All Figures

thumbnail Fig. 1

G31.581 + 00.077. a) Spitzer GLIMPSE 4.5–3.6 μm excess image overlaid with the water maser positions (yellow asterisks) from the EVN observation and H II region (yellow cross) detected by Bartkiewicz et al. (2009). b) Shows the distribution of water masers observed with the EVN (filled squares) from this paper and the VLA (open squares) from Bartkiewicz et al. (2011). The colours of the symbols relate to the LSR velocities as indicated on the right-hand side of the plot. The origin of the map is the position of the brightest 6.7 GHz methanol maser spot (Bartkiewicz et al. 2009) (see also Table 1). c) Shows an enlargement of the north-western (NW) water masers. d) Shows an enlargement of the south-eastern (SE) water masers together with the distribution of the 6.7 GHz methanol masers, marked by open circles, from Bartkiewicz et al. (2009). In c) and d), the arrows represent the velocity gradients (from blue- to red-shifted LSR velocities) detected within individual cloud. The black numbers and letters correspond to the clouds with internal velocity gradients, while the grey ones correspond to the clouds without internal velocity gradients (Table 3).

Open with DEXTER
In the text
thumbnail Fig. 2

Individual component spectra with Gaussian velocity profiles of 22.2 GHz water masers (red squares and lines) and 6.7 GHz methanol masers (blue circles and lines) towards G31.581 + 00.077 detected using the EVN in 2010 and 2007, respectively. The numbers and letters correspond to the cloud labels as given in Table 1 and Fig. 1. The grey lines show individual Gaussian profiles fitted to the blended features. The clouds with only two single spots are not marked to improve the clarity of the figure (i.e., clouds b, e, g).

Open with DEXTER
In the text
thumbnail Fig. 3

Same as Fig. 1 but for G33.641 − 00.228.

Open with DEXTER
In the text
thumbnail Fig. 4

Same as Fig. 2 but for G33.641−00.228. The scale on the left side describes the water maser intensity, while the right side corresponds to the methanol maser intensity. The clouds with only two single spots or without Gaussian characteristic are not marked for the clarity of the figure (i.e., clouds 5, 6, e, g, i, n, p).

Open with DEXTER
In the text
thumbnail Fig. 5

Velocity of the methanol maser spots (open circles) in G31.581 + 00.077 and G33.641 + 00.228 versus azimuth angle measured from the major axis. The sinusoidal line represents the best-fit kinematical model of a rotating and expanding disc of infinitesimal thickness to the methanol spots with the parameters listed in Table 4. For completeness, the water maser spots are shown as squares. In a case of G33.641 + 00.228 two kinematical models are presented for a powering source located as follow: D1 – in the centre of the best-fitted ellipse, D2 – at the MIR position.

Open with DEXTER
In the text
thumbnail Fig. 6

Outflow models fitted to water (squares) in G31.581 + 00.077 and G33.641 + 00.228 according to the model of Moscadelli et al. (2000). The relevant parameters are listed in Table 5. The right panel presents a comparison of obtained data (squares) vs. model (crosses). Vc is the systemic LSR velocity as given in Sect. 3.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.