Datasets:

SouthernCrossAI
/

Arxiv

Dataset card Data Studio Files Files and versions Community

Split (1)

train · 132k rows

Filename stringlengths 22 64	Paragraph stringlengths 8 5.57k
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The details of the physical processes that govern the earliest stages of high-mass star formation remain an area of intense research and much debate (e.g. Tan et al. 2014). Class II methanol masers are widely recognised as one of the most reliable tracers of the earliest stages of high-mass star forma- tion (Ellingsen 2006) and have the advantage that they are exclusively associated with this phenomenon (Breen et al. 2013).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The rich microwave spectrum of the methanol molecule contributes to the large number of diﬀerent transitions which show maser emission. Methanol masers are em- pirically divided into two classes, known as class I and class II (Menten 1991b). The population inversion for class I methanol masers is due to collisions(Cragg et al.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	limited sensitive and unbiased searches for class I methanol masers undertaken to date (e.g. Jordan et al. 2015, 2017; Yusef-Zadeh et al. 2013), suggest that the ATCA Legacy survey of the Galactic Plane at 7mm which commenced in 2017 will more than double the number of known sources over the next few years.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Infrared radiation creates the population inversion in class II methanol masers (e.g. Sobolev et al. 1997b; Cragg et al. 2005), with the 6.7- and 12.2-GHz transitions the most commonly observed and generally strongest. The majority of studies to date have focused on the class II methanol masers, as the strong transitions are at lower frequencies, (hence more readily observed) and early tar- geted searches towards other maser transitions and in- frared sources were very successful (e.g. Menten 1991a; MacLeod & Gaylard 1992; Caswell et al. 1995; Walsh et al. 1998). There are now more than 1000 class II methanol maser sites known in the Milky Way, with a number of large-scale, complete searches having been undertaken (e.g. Ellingsen et al. 1996b; Pandian et al. 2007), of which the Methanol Multibeam (MMB) is the most comprehensive (Green et al. 2010; Caswell et al. 2010, 2011; Green et al. 2012; Breen et al. 2015).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	A total of 18 diﬀerent class II methanol maser transi- tions or transition series have been observed toward high- mass star formation regions (see Ellingsen et al. 2012b, and references therein). The majority of these transitions have only been detected in a small number of sources and have signiﬁcantly weaker emission than is observed from the 6.7-GHz transition. Targeted searches have been made at 12.2-GHz towards all 6.7-GHz methanol masers detected in the MMB (Breen et al. 2012b,a, 2014, 2016), with detections towards 45.3 per cent of the sources. How- ever, the only other class II methanol maser transitions for which there have been sensitive searches towards moder- ately large samples of sources are the 19.9-GHz transition (Ellingsen et al. 2004, 22 sources), the 23.1-GHz transition (Cragg et al. 2004, 50 sources), the 37.7-, 38.3- and 38.5- GHz transitions (Haschick et al. 1989; Ellingsen et al. 2011, 2013, 106 sources), the 85.5-, 86.6- and 86.9-GHz transitions (Ellingsen et al. 2003, 22 sources) and the 107- and 156.6- GHz transitions (Caswell et al. 2000, 80 sources). Where multiple class II methanol maser transitions are observed to be co-spatial there is the potential to use them as probes of the physical conditions within the star formation re- gions (e.g. Cragg et al. 2001; Sutton et al. 2001). However, for many of the maser transitions observations have only been made with single-dish telescopes with angular resolu- tions of order arcminutes. Interferometric observations are required to determine whether the diﬀerent transitions are truly co-spatial, and hence whether the underlying assump- tion of the multi-transition maser modelling is valid (e.g. Krishnan et al. 2013). Here we present the ﬁrst interfero- metric observations of the 37.7-, 38.3- and 38.5-GHz class II methanol maser transitions. All of the sources in our sample have interferometric observations at 6.7-GHz available in the published literature, and many have also been observed at arcsecond angular resolution in the 12.2- and 19.9-GHz tran- sitions, enabling us to test the degree to which the emission is co-spatial over these transitions.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The observations were undertaken with the Australia Tele- scope Compact Array (ATCA) radio telescope on 2011 March 24 (project code C2191). The target sources were the eleven 37.7-GHz methanol masers from Ellingsen et al. (2011) within the declination range of the ATCA. The ar- ray was in the EW367 conﬁguration and the synthesised beam width for the observations at 38 GHz was approxi- mately 3 arcseconds × 5 arcseconds (although it was signiﬁ- cantly more elongated in declination for two near-equatorial sources). The EW367 array has baseline lengths for the in- ner ﬁve antennas ranging from 46 to 367 metres and the baselines from those antennas to sixth antenna are 4.0 – 4.4 km. In general only the short baselines are used from observations made in this array conﬁguration and this is the case for the majority of the investigations undertaken here, although we do use it to place limits on the brightness tem- perature. The astrometric accuracy of ATCA observations is primarily determined by the atmospheric conditions and the properties of the phase calibrators. The weather during the observations was excellent with the array seeing moni- tor reporting RMS path length noise of ∼ 70 µm and the astrometric accuracy is estimated to be approximately 0.4 arcseconds (see Section 3.1).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The Compact Array Broadband Backend (CABB) was used for the observations (Wilson et al. 2011) and conﬁg- ured with two 2 GHz bandwidths, each with a total of 32 × 64 MHz channels. At the time of the observations a single 64 MHz zoom band with 2048 spectral channels was available for each 2 GHz IF. For one of the IFs we alternated the zoom band (and IF) centre frequency between the 37710 MHz and 36175 MHz to cover the 37.7-GHz class II methanol maser transition and the 36.2-GHz class I methanol maser transition, respectively. For the other IF we alternated the zoom band (and IF) centre frequencies between 38300 MHz and 38460 MHz to cover the 38.3- and 38.5-GHz class II methanol maser transitions, respectively. Each zoom band has a total velocity coverage of ∼ 500 km s−1 and a veloc- ity resolution of 0.3 km s−1 for a uniform weighting of the correlation function.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	image was <∼ 25 mJy beam−1. We adopted rest frequencies (uncertainty), of 36.169290(0.000014), 37.703696(0.000013), 38.293292(0.000014) and 38.452653(0.000014) GHz for the 4−1 → 30E, 7−2 → 8−1E, 62 → 53A− and 62 → 53A+ transitions respectively (Xu & Lovas 1997). The uncertain- ties in the rest frequencies are all similar and correspond to an uncertainty in velocity of approximately 0.11 km s−1.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	For each transition where emission was detected the ab- solute position was measured by imaging a spectral chan- nel (usually of width 0.25 km s−1, except for the weakest sources) and ﬁtting the synthesised beam proﬁle to the im- age data. For sources with a peak ﬂux density stronger than approximately 300 mJy we then undertook a single itera- tion of phase self-calibration using a model created from the clean components of the initial imaging. In all cases this im- proved the signal-to-noise ratio in the image cubes, although in most cases the improvement was modest, consistent with good a priori calibration and the excellent weather condi- tions under which the observations were made. To image any radio continuum emission in each ﬁeld we combined the line-free data from each of the four zoom bands. To increase the sensitivity of the continuum images to large-scale struc- ture we used miriad’s multi-frequency synthesis option in the invert task and set Brigg’s robustness parameter to 1. We have not made any primary beam corrections in the current observations, as the majority of the emission from all transitions is in the central region of the primary beam where this is not necessary. Furthermore, the accuracy of the primary beam corrections depends critically on having a good model of the beam, which has recently been found to not be the case for the ATCA at 7-mm.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The sample for our study were the eleven 7−2 → 8−1E (37.7- GHz) methanol maser sources south of declination +5◦ ob- served by Ellingsen et al. (2011). Table 2 summarises the properties of the ten 37.7-GHz methanol masers detected in the current observations and the spectra extracted from the self-calibrated image cubes are shown in Figure 1. The velocity ranges given in Table 2 are the range over which the image cube shows emission present at the maser loca- tion with an intensity of 4-σ or greater (∼ 100 mJy). The spectra shown in Fig. 1 cover the entire velocity range of the 6.7-GHz methanol masers in each of the sources. Due to the large diﬀerence between the peak intensity and the weaker emission in many sources we have included scaled versions of the spectra of stronger sources in Figure 1 to facilitate investigation of all maser components.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	There are ten 37.7-GHz methanol masers detected in the current observations that were also observed by Ellingsen et al. (2011) and in addition the 38.3- and 38.5- GHz masers towards G 345.010+1.792 and NGC 6334F were also detected in those data. This gives a total of 14 spectra that can be compared between the current observations and those of Ellingsen et al.. We ﬁnd that for 9 of the 14 cases the current observations have weaker emission. Excluding the sources less than 5 Jy, for which the noise in the Mopra observations makes the amplitude measurements unreliable, the median across the sample shows the ﬂux density mea- sured in the current ATCA observations is 0.93 times that reported in the Mopra observations made two years previ- ous. This is consistent with the estimated accuracy of the Mopra and ATCA ﬂux density scales (15 per cent and 10 per cent respectively).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Some of the sources have multiple spectral features which show diﬀerences in the relative intensity between the Mopra and ATCA observations (e.g. G 323.740−0.263) and this, combined with the non-detection of G 35.201−1.736 demonstrates that at least in some sources the 37.7-GHz transition can show signiﬁcant temporal variability. If we exclude the 37.7-GHz methanol masers with a peak ﬂux den- sity < 5 Jy from consideration (as these are most susceptible to the eﬀects of noise), the ratios between the Mopra and ATCA peak ﬂux densities range from ∼0.6 – 1.3. So our re- sults are consistent with modest levels of variability being present in some 37.7-GHz methanol masers on timescales of 2 years.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Table 2 also gives the oﬀset between the absolute posi- tion of the 6.7-GHz methanol maser emission determined by Caswell (2009) and the current 37.7-GHz methanol maser observations (ten sources). The oﬀsets range from 0.3 – 1.5 arcseconds, with eight of the ten being less than 1 arc- second. In most cases the absolute positions of the two tran- sitions have been determined from spectral peaks with sim- ilar velocities, however, even where that is not the case, the strong 6.7-GHz methanol maser emission is generally conﬁned to regions with an angular extent of around 0.2 – 0.3 arcseconds (e.g. Phillips et al. 1998; Walsh et al. 1998), which will not produce a signiﬁcant oﬀset. The astromet- ric accuracy of ATCA observations with high signal-to-noise and good calibration is ∼0.4 arcseconds (Caswell 1997). The current observations employed four diﬀerent phase calibra- tors to accommodate target sources across a range of Galac- tic Longitudes, we have imaged each phase calibrator us- ing the adjacent calibrator and measured the position and ﬁnd the mean oﬀset to be 0.4 arceconds, consistent with ex- pectations. The mean oﬀset between the 6.7- and 37.7-GHz methanol masers is 0.68 arcseconds, with a median of 0.58 arcseconds. This diﬀerence is consistent with no oﬀset be- tween two independent observations each with an astromet- ric uncertainty of around 0.4 arcseconds. Hence we conclude that to within the astrometric uncertainty of the current observations the 6.7- and 37.7-GHz methanol masers arise from the same location.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Table 1. The coordinates of the pointing centre for each of the class II methanol maser sites observed. The RMS values refer to images made without any self-calibration.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Table 2. The properties of the 37.7-GHz methanol maser emission detected. The RMS values refer to the image data from which the spectra have been extracted.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The 36.2-GHz class I methanol maser transition was in- cluded in the observations because it could be observed si- multaneously with the 38.5-GHz transition with no nega- tive impact on the quality of the 37.7-GHz data collected. Somewhat surprisingly, 36.2-GHz class I methanol maser emission was detected towards all eleven sources observed. The properties of the detected maser emission are sum- marised in Table 5. The listed peak ﬂux density and ve- locity are considered over all class I methanol maser regions for that source, whereas the integrated values are extracted from the spectra shown in Figure 3, which have been ex- tracted from the image cube over a region which encom- passes all class I methanol maser regions within the primary beam of each source. A search of the published literature shows that four of the sources have not previously been re- ported to have class I methanol maser emission in any tran- sition (G 337.705−0.053, G 340.785−0.096, G 348.703−1.043 & G 35.201−1.736). Five of the sources (G 318.948−0.196, G 323.740−0.263, G 339.884−1.259, G 345.010+1.792 & NGC6334F) have been previously observed at 36.2-GHz in the large survey of Voronkov et al. (2014). The remaining two sources (G 9.621+0.196 & G 23.440−0.182) have pre- vious detections of both 44- and 95-GHz class I methanol masers (Kurtz et al. 2004; Gan et al. 2013; Slysh et al. 1994; Val’tts et al. 2000), but not for the 36.2-GHz transition.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	sites and this causes artefacts in the vector averaged spectra shown in Figure A10. To avoid this in the spectrum shown in Figure 3 we have used the scalar-averaged spectrum from the uv-data for G 23.440−0.182. The spectra in Fig. 3 incor- porate emission from all the 36.2-GHz methanol maser sites in each source, with the exception of G 337.705−0.053 and NGC6334F, where there is one class I methanol maser site (labelled E and F respectively) that is oﬀset signiﬁcantly from the rest and these have been excluded from the spec- tra. These spectra should be similar to those obtained by single dish observations towards the same regions. For ve- locities where the range of two or more individual 36.2-GHz methanol maser sites within one source overlap the ﬂux den- sities in Fig. 3 will exceed those shown for individual sites in Appendix A. Table 5 summarises properties of the spectra shown in Fig. 3.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	For the ﬁve sources previously observed with the ATCA in the 36.2-GHz transition, the ﬂux densities shown in Fig- ures A1 – A11 are systematically lower for the current obser- vations than in the corresponding ﬁgure in Voronkov et al. (2014). This is thought to be due to a combination of coarser spectral resolution in the current observations, an update to the ﬂux density scale at 7-mm for the ATCA since the ob- servations of Voronkov et al. and diﬀerences in the way the spectra have been extracted from the image data - extrac- tion at a single point in the current observations, compared with extraction over a region by Voronkov et al. (2014).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	We combined the line-free data from all the zoom band observations to create a continuum dataset for each tar- get source (eﬀective total bandwidth of approximately 220 MHz). We detected continuum emission toward 10 of the 11 ﬁelds, the exception was G 318.948 − 0.196, for which we can set an upper limit on any 8-mm continuum emis- sion of 2 mJy beam−1 (the RMS noise level in the image is 0.24 mJy beam−1). The continuum emission detected to- ward each source is shown in Figures A2 – A11 (in black or white contours, depending on the background) and the prop- erties of the continuum emission are summarised in Table 6. For eight of the ten sources where continuum emission is ob- served the class II methanol masers are projected against the continuum emission, in most cases close to the peak. Table 6 gives the oﬀset of the 37.7-GHz methanol maser (or 6.7-GHz methanol maser for G 35.201−1.736) from the peak of the radio continuum emission. The two sources where contin- uum emission is detected, but it is not associated with the maser site are G 348.703−1.043 and G 35.201−1.736.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 3. Spectra of the 36.2-GHz class I methanol maser emission extracted from the image cube data, with the exception of G 23.440−0.182, for which the spectrum is from scalar averaged uv-data.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Table 3. The properties of the 38.3-GHz methanol transition observations. The sources where the names are given in italics are new detections of this transition from the current observations. The spectral channels used to measure the RMS were 0.25 km s−1 for most sources, but 0.5 km s−1 for those indicated with ∗ and 1.0 km s−1 for those indicated with †. The RMS values refer to the image data from which the spectra have been extracted. NOTE: For G 337.705−0.053 the absorption in this transition was too weak to reliably image, so the position is assumed to be the same as for the 38.5-GHz transition.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Table 6. The properties of the 8-mm radio continuum emission detected. The oﬀset is between the peak of the radio continuum and the 37.7-GHz methanol maser, except for G 35.201−1.736, where the oﬀset is with respect to the 6.7-GHz methanol maser.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure A2 shows the general environment of the G 323.740−0.263 star formation region, with the class II methanol masers associated with an infrared source with a strong excess in the 4.5-µm wavelength band, charac- teristic of the presence of shock-excited gas. There is a large, evolved Hii region to the north-west of the maser source and the strongest class I methanol maser emission is to the south-east, although there is also emission asso- ciated with the infrared source and close to the class II maser site. Voronkov et al. (2014) identiﬁed four diﬀerent class I methanol maser sites in this source, all of which showed emission from the 44-GHz transition, with B and D also showing emission at 36.2-GHz. The current obser- vations detect 36.2-GHz methanol masers towards all four of the Voronkov et al. sites, and also at a new location to the south, which we have labelled E. The new site is be- yond the half-power point of the ATCA primary beam and Figure A2 shows that it is associated with a diﬀerent AT- LASGAL dust clump to the other maser emission. Hence, it is likely an unrelated nearby site of class I methanol masers, G 323.734−0.273, not directly associated with the G 323.740−0.263 maser region.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Combining the line-free channels from all four zoom- bands we are able for the ﬁrst time to detect the radio continuum emission associated with the class II methanol masers (see black contours in Figure A2). The emission has a peak intensity of 2.4 mJy beam−1 and is centred at right s h , Declination −56◦30′50.7′′, ascension (J2000) 15 45.71 an oﬀset of 3 arcseconds to the south-east of the strongest class II methanol maser emission. Previous ATCA observa- tions at 8.6 and 1.6 GHz found upper limits of 0.2 mJy and 0.4 mJy, respectively on the radio continuum emission at those frequencies (see Voronkov et al. 2014).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The higher sensitivity of the ATCA observations shows the 37.7-GHz methanol masers contain at least two spectral components over a velocity range -56.0 – -52.3 km s−1. One unique aspect of this source is that we observe absorption in the 37.7-, 38.3- and 38.5-GHz transitions centred at a velocity of approximately -49 km s−1. This can be seen in the spectrum of G 337.705−0.053 in Figure 1, but is more readily apparent in the 38.3- and 38.5-GHz data due to the absence of emission in those transitions (see Fig. 2). Im- ages of the 38.5-GHz transition shows that the absorption is located in front of the radio continuum emission associ- ated with the class II methanol maser emission with the strongest absorption oﬀset approximately 1 arcsecond north of the continuum peak.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	any meaningful comparison of the relative distribution of the 6.7- and 37.7-GHz masers in this source.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure A3 shows the class II methanol maser emission and radio continuum lies at the edge of an extended infrared source, but close to the peak of the ATLASGAL 870-µm dust emission. The bulk of the 36.2-GHz class I methanol maser emission lies along the western-edge of the extended infrared source and the strongest emission is redshifted with respect to the 37.7-GHz methanol maser emission and the absorption seen in the other transitions.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The higher sensitivity of the ATCA observations shows that the 37.7-GHz methanol masers have a much larger veloc- ity range than was apparent from the single dish observa- tions of Ellingsen et al. (2011). G 339.884−1.259 remains the strongest 37.7-GHz methanol maser source in the Galaxy, with a peak ﬂux density of around 300 Jy. In addition to the strong component there are two weaker components each with a ﬂux density of approximately 1.5 Jy with peak ve- locities of -34.7 and -33.5 km s−1, which can be seen in the dashed spectrum for this source in Figure 1. Interestingly, G 339.884−1.259 has recently been discovered to also host the strongest 4.8-GHz formaldehyde maser in the Galaxy (Chen et al. 2017) and demonstrates a number of spectral components which closely align in velocity with 6.7-GHz methanol maser features.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The greater sensitivity has also resulted in the detection of weak maser emission in the 38.3- and 38.5-GHz transitions in this source (Fig. 2). The proﬁle of the two transitions is similar, with two components with peak velocities of ap- proximately -38.5 and -34.0 km s−1. The 36.2-GHz class I methanol maser emission in this source is relatively weak (peak ﬂux density < 2 Jy). Figure A4 shows the location of the class I methanol masers with respect to the other masers and continuum emission in the G 339.884−1.259 star formation region. The distribution matches that observed by Voronkov et al. (2014), but we detect additional weak 36.2-GHz emission (component C), between components A and B. The radio continuum emission and class II methanol masers coincide with the peak of the infrared emission seen in the WISE image and the ATLASGAL dust continuum emission shows elongation with approximately the same po- sition angle seen in the class II methanol maser cluster (see for example Dodson 2008) and on larger scales in the class I methanol maser emission.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	to make a detailed comparison with the complex spatial dis- tribution of the 6.7-GHz methanol masers seen in ATCA observations (Norris et al. 1993; Phillips et al. 1998).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	There is no 38.3- or 38.5-GHz methanol emission stronger than 100 mJy, however, we do detect 36.2-GHz class I methanol maser emission. Figure A5 shows the en- vironment of the G 340.785−0.096 region, with the class II methanol masers located at the peak of an ATLASGAL dust clump, which is located within a low-contrast infrared dark cloud. The class II methanol masers are close to the peak of the radio continuum emission which lies between two com- pact infrared sources. The class I methanol maser emission is complex with at least 6 diﬀerent sites distributed around the class II masers and radio continuum emission. The ma- jority of the class I methanol maser emission is the in the same velocity range as the class II methanol masers, how- ever, there are weak class I masers at site A, to the north-east of the class II masers which show emission in the velocity range -151.5 – -141.0. The most blue-shifted emission from any of the other sites is at -107.3 km s−1, so this compo- nent has a velocity oﬀset of > 30 km s−1. Voronkov et al. (2014) found that high-velocity class I methanol maser emission is relatively rare, with the majority of cases de- tected as blueshifted components in the 36.2-GHz transi- tion. Of the sources observed by Voronkov et al. (2014), G 340.785−0.096 appears most similar to G 309.38−0.13, as the high-velocity emission has a relatively small angular sep- aration from the other sites. The high-velocity 36.2-GHz class I methanol emission in G 309.38−0.13 has been studied in detail by Voronkov et al. (2010a) and subsequent obser- vations have detected 44-GHz methanol masers at the same location (Voronkov et al. 2014). Detailed investigations of regions such as this may help us gain insight into the more extreme class I methanol maser emission seen in the Galactic Centre (Yusef-Zadeh et al. 2013) and some starburst galax- ies (Ellingsen et al. 2014, 2017; McCarthy et al. 2017).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	detection, which we would not have reported without the in- dependent conﬁrmation from the previous observations.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	G 345.010+1.792 lies outside the regions covered by both the ATLASGAL (Schuller et al. 2009) and GLIMPSE (Benjamin et al. 2003; Churchwell et al. 2009) surveys, but has been imaged in other Spitzer observations. Figure A6 has a Spitzer three-colour image as the background (data extracted from the archive) and shows the class II methanol masers are projected against the radio continuum peak and are associated with a strong, extended infrared source. The class I methanol masers are located to the north-west of the class II site, with some of the emission projected against a region of infrared emission which has an excess in the 4.5-µm band, indicative of shock-excited gas.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The ATCA observations show a single, relatively weak 37.7- GHz class II methanol maser in this source with no accompa- nying emission in either the 38.3- or 38.5-GHz transitions. The newly detected 36.2-GHz class I methanol masers in this source are at the blueshifted end of the 6.7-GHz veloc- ity range, in contrast to the 37.7-GHz emission which lies near the redshifted end. Figure A7 shows that the class II methanol masers are oﬀset from the peak of the ATLASGAL dust clump by approximately 10 arcseconds, at the edge of a region of diﬀuse, weak infrared emission. The strongest class I methanol masers are closer to the peak of the dust continuum emission. There is no radio continuum emission associated with any of the methanol maser sites, but there is a strong radio continuum source oﬀset approximately 1 ar- cminute to the west of the masers (see Fig. A7). This radio continuum emission is associated an infrared source showing a strong excess in the 4.5-µm band.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	This is an iconic star formation region which has been widely studied, including the class II methanol maser emission. The 37.7-GHz methanol masers detected in the ATCA observa- tions show a strong peak towards the blueshifted end of the spectrum, with weaker emission extending to -2.8 km s−1, but no other distinct spectral peaks and no other emission stronger than 1 Jy (there is a small shoulder in the gradual declining strength of the emission at -8 km s−1). NGC6334F is unique in that the 38.3- and 38.5-GHz methanol masers are stronger than the 37.7-GHz masers. This was also the case in the observations of Ellingsen et al. (2011) and com- paring the spectra of the three transitions between the two epochs (separated by approximately two years) we can see that the 37.7-GHz emission has decreased to approximately 70 percent observed by Ellingsen et al. (2011), while there is little diﬀerence in the 38.3- and 38.5-GHz emission.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	towards both of these sites and this is discussed in more de- tail in section 4.1. The 36.2-GHz class I methanol masers in this source have previously been observed by Voronkov et al. (2014), who ﬁnd ﬁve class I maser sites, four of which (A, B, C and E) have 36.2.GHz masers. We detect 36.2-GHz emission towards three of the four sites - we do not ﬁnd any convincing emission at site C, which is the weakest of the 36.2-GHz maser locations in Voronkov et al. (2014). We also detect strong 36.2-GHz methanol maser emission from a site outside the primary beam of the ATCA to the north (la- belled F in Figure A8. This emission is associated with the well-studied class I methanol maser source NGC6334I(N) (e.g. Kogan & Slysh 1998; Beuther et al. 2005; Brogan et al. 2009).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	This is the strongest 6.7-GHz methanol maser source and also shows periodic variability in a number of class II methanol maser transitions (e.g. Goedhart et al. 2003). The ATCA observations conﬁrm the speculation of Ellingsen et al. (2011) that there is weaker 37.7-GHz maser emission covering the velocity of the strongest 6.7-GHz methanol masers (which peak at +1.3 km s−1). The cur- rent observations also detect very weak (∼0.2 Jy) emission from the 38.3-GHz transition, but not from the 38.5-GHz transition. Signiﬁcant diﬀerences in the ﬂux density of the 38.5-GHz masers compared to the 38.3-GHz masers has pre- viously only seen towards G 335.789+0.174, (Ellingsen et al. 2013).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure A9 shows that the class II methanol masers are located at the edge of an extended infrared source which also has associated radio continuum emission. The class II methanol maser site is close to several regions which show an infrared excess in the 4.5-µm wavelength range and the class I methanol maser emission is predominantly to the east of the infrared, radio continuum and class II methanol.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Fig. A10 due to interference between emission at diﬀerent locations in velocity ranges that overlap.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	As discussed in Section 3.1 we do not detect any 37.7-GHz methanol masers at the epoch of these observations with a 5-σ upper limit of 80 mJy. We also do not detect any emis- sion in either the 38.3- or 38.5-GHz transitions. We do detect 36.2-GHz class I methanol masers close to G 35.201−1.736, although Figure A11 shows that they are signiﬁcantly oﬀset from the class II methanol maser site and likely associated with a nearby, unrelated infrared source. G 35.201−1.736 lies outside the ATLASGAL and GLIMPSE survey regions and Fig. A11 shows a WISE three-colour infrared image as the background and there is strong infrared emission in the ﬁeld which has saturated some of the bands. G 35.201−1.736 is located only a little more than a degree from declination zero and as an east-west oriented array, the synthesised beam of the ATCA observations is nearly one-dimensional for this source. Because of this we have used a 4 arcsecond circular restoring beam for both the radio continuum and integrated 36.2-GHz methanol maser data presented in Fig. A11 and while the right ascension positions will be accurate, there is signiﬁcant uncertainty in the declinations compared to the other sources observed. Figure A11 suggests that the radio continuum peaks at the location of the infrared peak, with the class II methanol masers, which here the position of the 6.7-GHz class II methanol masers reported by Breen et al. (2015) are associated with an infrared point source with ex- cess emission in the WISE 4.6-µm band which lies at the edge of the strong extended infrared source. We identify three sites of class I methanol masers west of the class II masers and radio continuum emission, however, proper iden- tiﬁcation of the distribution of the class I masers will re- quire observations in an array conﬁguration with north- south baselines.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	where both transitions are observed the peak velocities are within 1 km s−1 (Breen et al. 2011). The demonstration that the two transitions are co-spatial in some sources, com- bined with the good agreement in the line of sight velocity of the peak emission in the majority, strongly suggests that the 6.7- and 12.2-GHz commonly arise from the same re- gion (to within spatial scales of around 0.05 pc). The degree to which this close association holds for the other class II methanol maser transitions remains an open question, but one which the current observations can help to address.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Krishnan et al. (2013) used the ATCA to make obser- vations of the 19.9-GHz (21 → 30E) transition towards 9 class II methanol maser sources (and the 23.1-GHz 92 → 101A+ towards one source). They found good agreement be- tween the absolute positions of the 6.7- and 19.9-GHz class II methanol maser sites (∼0.2 arcseconds), but did not ﬁnd any evidence for close correspondence between the maser spots from the two transitions. One of the complicating factors in their investigation was that the 19.9-GHz methanol masers had relatively few spectral components (hence few spots) and in particular there were few sources with multiple strong components. This makes comparison of the relative spatial distribution of the 19.9-GHz maser spots compared to the 6.7-GHz masers more diﬃcult.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	The 37.7-GHz methanol maser transition is typically the fourth strongest of the class II transitions (after the 6.7-, 12.2- and 107-GHz transitions) and hence potentially bet- ter for such comparisons. Figure 1 shows the spectra of the 37.7-GHz methanol masers and demonstrates that while the majority of the sources have multiple spectral components when observed with suﬃcient sensitivity, most have only a single strong component in the 37.7-GHz transition. In addi- tion, where there are multiple strong spectral components it is only possible to investigate the relative distribution when they have a spatial separation greater than around 0.1 arc- seconds. When measuring the position of the emission in each channel of the image cubes, we found that consecutive channels with emission stronger than approximately 1 Jy typically give good agreement (within a few 10s of milliarc- seconds), whereas for weaker emission the scatter is larger (0.1 – 1 arcsecond). This is consistent with the expectation that the accuracy to which the relative position of emission in diﬀerent spectral channels can be determined depends on the size of the synthesised beam and is inversely propor- tional to the signal-to-noise ratio. So for the purposes of making comparison of the 37.7-GHz methanol masers with the 6.7-GHz distribution we focus on the components with a ﬂux density greater than 1 Jy.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 4. The relative distribution of 6.7-GHz methanol masers determined by Walsh et al. (1998) (crosses) and 37.7- GHz methanol masers (squares) in G 339.884−1.259. For the 6.7- GHz methanol masers the size of the cross is proportional to the ﬂux density of the component, while for the 37.7-GHz methanol masers all components are shown at the same size, as the strongest (-38.7 km s−1) component is several orders of magnitude stronger than all the others.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Section 3.5 for further details), whereas for the other three sources there is some agreement.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 4 shows a comparison of the 6.7- and 37.7-GHz methanol masers in G 339.884−1.259. The absolute posi- tions have been shifted to better align the two transitions and we can see that the relative distribution is similar in terms of the orientation and velocity gradient of the emis- sion. The position angle of the elongation seen in the class II methanol masers in this ﬁgure agrees with that seen on much larger scales in Figure A4 in the class I methanol masers and dust continuum emission.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 5 shows a comparison of the 6.7- and 37.7-GHz methanol masers in G 345.010+1.792. Here we only detect two components at 37.7-GHz and the absolute positions have been shifted to align the -22.4-km s−1 component from the Norris et al. (1993) 6.7-GHz distribution with the -22.0- km s−1 component at 37.7-GHz. We can see that when we do that the secondary component at 37.7-GHz is oﬀset in the same general direction and sense of velocity gradient as observed at 6.7-GHz, but that there is a diﬀerence in the ve- locity of the 37.7-GHz secondary component and the nearest 6.7-GHz counterpart.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 5. The relative distribution of 6.7-GHz methanol masers determined by Norris et al. (1993) (crosses) and 37.7-GHz methanol masers (squares) in G 345.010+1.792.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 6. The relative distribution of 6.7-GHz methanol masers determined by Norris et al. (1993) (crosses) with the 37.7-GHz methanol masers (squares) and 38.3-GHz methanol masers (cir- cles) in NGC6334F. The 37.7- and 38.3-GHz positions have been determined from ATCA data including all antennas, which gives a synthesised beam of approximately 2.3 × 1.4 arcseconds.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	north-east of the northern-most 6.7-GHz site. To minimise this issue, we imaged the 37.7-, 38.3- and 38.5-GHz methanol masers in NGC6334F using the data from all ATCA an- tennas which yields a synthesised beam of approximately 2.3 × 1.4 arcseconds. These data have a large gap in the uv-coverage, however, because the morphology of the maser emission is point-like we are still able to obtain reasonable images. At this higher angular resolution the 37.7-, 38.3- and 38.5-GHz methanol masers can clearly be seen to origi- nate in two clusters, associated with the same regions which show the strongest 6.7- and 12.2-GHz methanol masers in Figure 6.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 7 shows the location and extension of the in- tegrated 37.7-GHz methanol maser emission with respect to the radio continuum, infrared emission and location of millimetre continuum sources. The two main sites of 6.7- GHz methanol masers are towards the continuum sources labelled as MM3 (the well known ultra-compact Hii re- gion which which has associated radio continuum emission) and MM2 (which does not show centimetre radio contin- uum) (Brogan et al. 2016). Figure 6 shows that the 38.3- GHz maser distribution associated with the MM3 region extends north-west from that observed at 6.7-GHz and for the emission from the MM2 region it is displaced to the east. The 37.7-GHz masers show the same trend, but to a lesser extent. This is almost certainly due to small degree of residual spectral/spatial blending between diﬀerent sites for these transitions. Unfortunately, for the 37.7-, 38.3- and 38.5-GHz transitions there is only a single strong maser com- ponent, with all the emission towards the MM2 site being much weaker (around 1 Jy at 37.7-GHz and < 0.5 Jy in the other two transitions). This means that it is not possible with the current data to rigorously test whether the relative distributions of any of these class II maser transitions show close correspondence with the 6.7- and 12.2-GHz masers. We also imaged the 38.5-GHz methanol masers at higher angu- lar resolution, the results are very similar to those observed for the 38.3-GHz transition. To avoid clutter and confusion within the images we show only the 38.3-GHz results, but the comments below relating to the spatial distribution of the 38.3-GHz methanol masers in NGC6334F apply equally to the 38.5-GHz transition.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	emission with ﬂux density of approximately 2 Jy, present at redshifted velocities (around -8 km s−1) and weaker emis- sion in the same velocity range at 38.3-GHz. While towards MM1 there seems to be a broad, weak spectral component present in both transitions. This strongly supports the hy- pothesis that the 37.7-, 38.3- and 38.5-GHz methanol emis- sion is located at three sites within NGC6334F, associated with MM3, MM2 and MM1. We would expect that observa- tions with higher angular resolution (0.5 arcsecond or bet- ter) would localise the 37.7- and 38.3-GHz emission towards MM3 and MM2 and show better agreement with the 6.7- and 12.2-GHz results. The emission associated with MM1 may be thermal, or quasi-thermal emission, rather than maser emission and so may be resolved out in higher resolution observations. The 19.9- and 23.1-GHz class II transitions in NGC6334F have also been imaged using the ATCA, how- ever, for those transitions emission is only observed towards the MM3 site (Krishnan et al. 2013).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	In summary, from our imaging data we observe good agreement in the absolute positions (i.e. within the expected astrometric accuracy), and a broad similarity in the distri- bution (for three of the four sources where this can be inves- tigated). This is good evidence that in general the 37.7-GHz methanol masers arise from the same sites as the 6.7- and 12.2-GHz masers (and likely all the other class II transi- tions). However, with the current data we are not able to test whether some of the individual 37.7-GHz maser spots are co-spatial with those at 6.7-GHz and 12.2-GHz. For most sources this is because there are not enough strong spec- tral components in the 37.7-GHz emission to enable rigor- ous comparison with other class II maser distributions. The best prospect for a more deﬁnitive test appears to be higher angular resolution observations of the 37.7-GHz methanol masers in the NGC6334F star formation region, as emission is observed over a wide velocity range at multiple sites.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	One additional test that can be made with the cur- rent data is to make a detailed comparison of the spectra of the 6.7- and 37.7-GHz masers. To demonstrate that emis- sion is co-spatial between two transitions it is necessary to show both that the emission arises from the same location at high angular (spatial) resolution, and also to demonstrate agreement in the spectral characteristics (peak velocity, line shape etc). The individual maser spots arise when there is a suﬃciently long path through the region of masing gas, that has a high degree of line-of-sight velocity coherence in the speciﬁc direction of the observer. This does not require the entire path to share a speciﬁc line-of-sight velocity to the observer, just a suﬃcient fraction of that particular path. The physical conditions within the region where the masers arise will not be completely uniform, so since the various maser transitions have diﬀerent sensitivities with respect to changes in the physical conditions we expect diﬀerent parts of the velocity coherent line-of-sight to the observer will favour diﬀerent transitions. The net result of this is that two transitions can have co-spatial maser spots in the plane of the sky, but still exhibit diﬀerences in the line pro- ﬁle due to diﬀerences in the line-of-sight contributions. This will always be the case to some degree, however, similarities (and diﬀerences) in the spectral proﬁles provide some inde- pendent information as to the degree to which the emission might be co-spatial between transitions.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	same peak velocity to within 1 km s−1 in 78 percent of sources (within 0.2 km s−1in 56 percent) is consistent with a high degree of co-spatial emission between the two tran- sitions (Breen et al. 2011). Comparing the peak velocity of the 37.7- and 6.7-GHz methanol masers we ﬁnd that they have the same peak velocity to within 1 km s−1 in 7 of 10 sources, comparable to the 6.7- and 12.2-GHz result. Only 3 of 10 sources have the peak velocities agreeing to within 0.2 km s−1, but for such a small sample this is not sig- niﬁcantly diﬀerent to the 6.7- and 12.2-GHz results. Fig- ure 8 shows a comparison between the 37.7-GHz and 6.7- GHz methanol maser emission in the ten sources detected in the current observations. The 6.7-GHz methanol maser data are the so-called MX spectra from the methanol multi- beam survey (MMB), which were observed with the Parkes telescope in 2008 and 2009 (Green et al. 2010; Caswell et al. 2010, 2011; Green et al. 2012; Breen et al. 2015). To facil- itate the comparison between the spectral components the 6.7-GHz methanol maser spectra have been scaled so that it has the same peak ﬂux density as the 37.7-GHz methanol emission. The 37.7-GHz methanol maser data is that ex- tracted from the image cubes, which have a 0.25 km s−1 velocity resolution (note that the intrinsic resolution of the observations is slightly coarser than this at 0.3 km s−1). The 6.7-GHz methanol maser spectra have a spectral resolution of 0.11 km s−1.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 7. Top Panel: The background image is from Spitzer IRAC observations from the GLIMPSE survey (Benjamin et al. 2003) of the NGC6334F region with blue, green, and red from the 3.6, 4.5, and 8.0 µm bands, respectively. The black dashed contours are the 8-mm radio continuum emission from the current ATCA observations (inner 5 antennas), with contours at 2, 4, 8, 16, 32 and 64 and 90 percent of 2.01 Jy beam−1. The red contours are the integrated 37.7-GHz methanol maser emission utilising data from all ATCA antennas with contours at 2, 4, 8, 16, 32 and 64 percent of 29.4 Jy km s−1 beam−1. The blue stars mark the locations of the millimetre sources MM1, MM2 and MM3, (from north to south, respectively), identiﬁed by Brogan et al. (2016). The spectra to the left of the image show the 37.7-GHz methanol maser emission extracted from the image cube at the speciﬁed locations and the spectra to the right show the 38.3-GHz methanol maser emission from the same locations. The red-ﬁlled ellipse in the bottom-left corner represents the synthesised beam of the 37.7-GHz observations. The x-axis of the spectra are line-of-sight velocity in km s−1 and the y-axis of the spectra show the ﬂux density in Jy. Bottom Panel: The region of the top image marked by the yellow-dashed square. The black-dashed and red contours and the blue stars are the same as in the top image, the squares and circles show the distribution of the 37.7- and 38.3-GHz methanol masers, respectively, from Fig. 6.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	primarily observed in 2008 and 2009, several years prior to the 37.7-GHz observations, as were the 12.2-GHz methanol maser follow-up observations of Breen et al.. So variability can potentially aﬀect the comparison of the spectra of diﬀer- ent transitions toward the same source, however, the MMB observations and comparison with earlier data shows that peak velocity of strong 6.7-GHz methanol masers generally does not change greatly on timescales of a few years. So it is unlikely that variability is playing a signiﬁcant role in the diﬀerences observed between the transitions. Combining our investigation of the spatial distribution of the 37.7-GHz methanol masers and the spectral comparison we conclude that while the 37.7-GHz methanol masers may arise from the same sites as the 6.7-GHz methanol masers they are less often, or to a lesser degree co-spatial than the 12.2- GHz masers are with the 6.7-GHz transition. We note that in agreement with theoretical predictions, our observations suggest that the brightness temperatures of the 37.7-GHz masers are always lower than those of the 6.7-GHz and 12.2- GHz methanol masers at the same velocity. However, sensi- tive observations made at the same epoch with higher spec- tral resolution (particularly for the 37.7-GHz spectra) are required to provide deﬁnitive data for testing the degree of co-spatiality and relative brightness temperatures.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	the MMB survey for the 6.7-GHz masers associated with the ten 37.7-GHz detections from the current observations, we ﬁnd that 70 percent are in the blueshifted half of the veloc- ity range (Figure 9, right-hand panel). This suggests that the blueshifted tendency for 37.7-GHz methanol masers is signiﬁcantly stronger than that for the larger population of class II methanol masers.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	If we consider the location of the 37.7-GHz methanol maser peak velocity within the 37.7-GHz velocity range the results are more extreme, with 8 of the 10 sources within the blueshifted half of the velocity range. Furthermore, 0 of the 10 sources are in the most redshifted 30 percent of the range and 6 of the 10 sources in the most blueshifted 30 percent of the range. If we assume that the location of the peak velocity of the 37.7-GHz methanol masers within the velocity range is uniform, then the probability of 6 or more from a sample of 10 sources being within the most blueshifted 30 percent is less than 5 percent, while the chance of 0 from 10 being within the redshifted range is less than 3 percent. So it is unlikely that the peak of the 37.7-GHz methanol masers has an equal chance of being anywhere within the velocity range of the emission, our data provide good evidence that the peak is preferentially towards the blueshifted half of the range. Furthermore, the class II methanol masers with an associated 37.7-GHz methanol maser are more likely to have a peak in the blueshifted half of the velocity range than the overall population.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	How does the class II methanol maser peak and veloc- ity range compare or relate to the general molecular gas in a region? A number of studies have shown that both the peak and mid-range velocity of class I methanol maser sources are an excellent proxy for the systemic velocity of the region (e.g. Voronkov et al. 2014; Jordan et al. 2017; Yang et al. 2017). We have made a comparison between the peak velocity of the 37.7-GHz class II methanol masers and the 36.2-GHz class I methanol masers and ﬁnd that for eight of the ten sources from the current sample, the class II methanol maser peak is blueshifted compared to the class I methanol maser peak. We performed a similar analysis comparing the veloc- ity of the 36.2-GHz methanol maser peak for each source observed by Voronkov et al. (2014), with the velocity of the 6.7-GHz methanol maser peak measured in the MMB sur- vey. Voronkov et al. (2014) observed 71 36.2-GHz methanol maser sources and 62 of these have a 6.7-GHz methanol maser within 1 arcminute (where there was more than one 6.7 GHz methanol masers with 1 arcminute, the source with the smallest angular separation was assumed to be associ- ated). For this sample of 62 sources 39 (62 percent) have the class II methanol maser emission redshifted with respect to the class I methanol maser peak velocity. So once again, the 37.7-GHz methanol maser sources show very diﬀerent behaviour to larger, more representative samples. The ten- dency for the 37.7-GHz methanol maser peak velocity in our observations to be blueshifted compared to the systemic velocity suggests that they are preferentially observed from molecular gas outﬂowing with respect to the systematic ve- locity and directed towards the observer.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 8. Spectra of the 37.7-GHz methanol maser emission (solid line), compared to the 6.7-GHz methanol masers scaled to have the same peak ﬂux density as the 37.7-GHz emission (red dashed line).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Figure 9. The location of the peak of the 6.7-GHz methanol maser emission as a fraction of the total velocity range of the 6.7-GHz emission. The left-hand plot shows the distribution for the 6.7-GHz methanol masers in the Caswell et al. (2011) paper (198 sources). The right-hand plot shows the same information (location of the 6.7-GHz methanol maser peak within the 6.7-GHz velocity range) for the 37.7-GHz masers.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	and 38.5-GHz class II methanol maser transitions. Although we have only used the data for the inner-ﬁve antennas for the majority of the analysis, the data to the sixth antenna was recorded. Analysis of the longer baseline data for the strongest component in all the masers with a peak ﬂux den- sity greater than 5 Jy (i.e. 7 of the 10 sources at 37.7-GHz and 2 sources at 38.3-/38.5-GHz) shows that in all cases the emission is unresolved on the longest baselines (approxi- mately 4.4 km). This corresponds to a uv-distance of approx- imately 540 kλ, or an angular scale of ∼0.38 arcseconds.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	respectively. Brightness temperatures in excess of 1011 K are observed for some 6.7- and 12.2-GHz methanol masers (e.g. Menten et al. 1988, 1992).
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	observations show that the ratio of the lower limit of the 6.7-GHz masers (> 1011.5 K) to the lower limit of the 37.7-, 38.3- and 38.5- GHz methanol masers to be 10−5.3, 10−6.6 and 10−6.8 respectively for G 345.010+1.792, quite diﬀerent from what is observed for NGC6334F.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	A focus of a number of diﬀerent large-scale maser surveys has been to use the results to established a statistically- based evolutionary timeline for masers associated with high-mass star formation regions (e.g. Breen et al. 2011; Voronkov et al. 2014; Titmarsh et al. 2016; Jordan et al. 2017). These investigations are aiming to quantify the qualitative evolutionary timelines that have been proposed (Ellingsen et al. 2007; Breen et al. 2010). Where rare, weak maser transitions such as the 37.7-GHz methanol masers ﬁt within such schemes has been discussed in some de- tail previously and we will not repeat that discussion here (Ellingsen et al. 2011, 2013; Krishnan et al. 2013). Ellingsen et al. showed that 37.7-GHz methanol masers are only found in star formation regions with the highest 6.7- and 12.2-GHz methanol maser luminosities and suggested that that indicated they arise towards the end of the class II methanol maser phase.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Here we investigate the new information provided by the more sensitive high-resolution observations and its im- plications for the evolutionary timeline. Some of the anal- ysis undertaken requires the distance to the source to be known or estimated and we have tabulated this information in last column of Table 5. For 5 of the 11 target sources we have trigonometric parallax measurements of the distance to the sources, for the remaining 6 sources we have used the Baysian distance calculator of Reid et al. (2016), assuming the near kinematic distance when that is the primary means by which the distances have been inferred. For some of the sources the kinematic distance estimates are a little diﬀer- ent from those used in previous works, as we have used the peak velocity of the 36.2-GHz methanol masers as the line- of-sight velocity, as that is a more reliable estimate of the systemic velocity (see discussion in Section 4.2), however, it diﬀers from the class II methanol maser peak (and mid- point) velocity.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	MM3, which has a strong ultracompact Hii region and OH masers (Forster & Caswell 1999; Brooks & Whiteoak 2001). The two class II methanol maser sites have similar millime- tre ﬂux densities, but Hunter et al. (2006) estimate that MM3 is the more massive young stellar object. Overall, Hunter et al. (2006) ﬁnd 4 high-mass young stellar objects in the NGC6334F region and one possibility for the pres- ence of multiple sites of 37.7-GHz methanol maser emission is that in dense clusters it can arise due to evolution of the gas and dust in the system, rather than that associated with a single young stellar object.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	Voronkov et al. (2014) found that the linear spread of class I methanol maser features and the total veloc- ity width of the class I emission was signiﬁcantly larger for those which had an associated OH maser compared to those which did not (see ﬁgure 5 of their paper). In their sample, the class I methanol masers without an as- sociated OH maser all had spatial extents < 0.4 pc and velocity ranges < 10 km s−1, whereas the majority of OH- associated sources exceeded one or both of these limits. All of the eleven target sources for the current observations have an associated ground-state OH maser (see table 4 of Ellingsen et al. 2013), however, we exclude G 35.201 − 1.736 from consideration, as the class I methanol masers don’t appear to be associated with the class II masers (see com- ments for this source in Section 3.5). We measured the linear spread of the class I masers by ﬁnding the angular sepa- ration between the most widely separated 36.2-GHz com- ponents for each source, using the positions recorded in Table A1 and converting to a linear scale using the dis- tance estimates from Table 5. We ﬁnd four of the sources (G 318.948−0.196, G 339.884−1.259, G 340.785−0.096 and G 23.440−0.182) have a linear spread >0.4 pc, with a further two sources having velocity widths greater than 10 km s−1 (G 337.705 − 0.053 and G 9.621 + 0.196). So in total 6 of 10 sources have either a linear spread >0.4 pc (4 sources), a velocity width >10 km s−1 (4 sources), or both (2 sources). It should be noted that 5 of the 11 sources here are in- cluded in the Voronkov et al. (2014) sample (and are all OH- associated sources) and that observations of Voronkov et al. included both the 36.2- and 44-GHz methanol maser transi- tions and the latter of these is generally more common and stronger in high-mass star formation regions. This means that the linear extents measured in the current observations are likely systematically lower than those in Voronkov et al. (2014) and hence the current observations are consistent with the class I methanol masers being associated with more evolved star formation regions.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	to the end of the class II methanol maser stage, is further evidence that some class I methanol maser are associated with older more evolved star formation regions. The ﬁgures in Appendix A show that where there is continuum emission associated with the class II methanol maser site, the class I masers are typically distributed around it, again suggesting an association with a more evolved star formation region. However, it should be noted that sensitive observations, such as those reported here, detect class I methanol masers to- wards a very high percentage of high-mass star formation regions. The MALT-45 survey of Jordan et al. (2017) made a blind survey of a 5 square degree region of the Galactic Plane which detected 77 class I methanol maser sources (in the 44-GHz transition). Of these 77 masers, 31 are associ- ated with a 6.7-GHz class II methanol maser, showing that more than half of the class I maser sources are not associ- ated with a class II maser. Jordan et al. (2017) ﬁnd that the majority of sources have a linear extent <0.5 pc and a ve- locity width <5 km s−1 and also that those class I methanol masers which are not associated with OH masers or radio recombination lines are generally less luminous than those which are. The class I methanol masers are typically associ- ated with dust clumps with masses > 1000M⊙, suggesting the majority are associated with high-mass star formation regions. From this they infer that the majority of the class I methanol masers without an associated class II methanol maser are younger sources.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	dance (noting that such shocks themselves tend to enhance methanol abundance through heating and evaporation of ices from dust grain mantles). In this scenario the class I methanol maser phase traces the duration over which a star formation region has suﬃciently high methanol abundance - from the release of methanol into the gas-phase from icy grain mantles, to the eventual destruction of the gas-phase methanol through a combination of chemical processing and dissociation from high-energy radiation or shocks. Whereas the class II methanol maser phase requires both the presence of gas-phase methanol and appropriate physical conditions (temperature, density etc) in the gas close to the young star. Hence, we expect the radiatively-pumped OH and class II methanol masers to be much more directly inﬂuenced by the young high-mass star and hence a better tracer of evolution- ary phase than the collisionally pumped water and class I methanol masers.
Processed_The_first_high-resolution_observations_of_37.7-,_3.txt	We have made the ﬁrst high resolution observations of the 37.7-, 38.3- and 38.5-GHz class II methanol maser transi- tions towards a sample of eleven sources. We ﬁnd that to within the absolute astrometric accuracy (around 0.4 arcsec- onds), these class II transitions arise from the same location as the strong 6.7-GHz class II methanol masers. The spectral and imaging data both suggest that the 37.7-GHz masers are less co-spatial with the 6.7-GHz maser spots than has pre- viously been observed for the 12.2-GHz class II transition. The peak velocity of the 37.7-GHz masers is found to be pref- erentially blueshifted with respect to the systemic velocity, suggesting that these maser arise in molecular gas which is outﬂowing from the region in the general direction of the observer. The detection of class I methanol masers towards what are hypothesised to be star formation regions towards the end of the class II methanol maser phase, combined with the results of recent untargetted searches for class I methanol masers suggest that the duration of the class I methanol maser phase exceeds that of class II methanol masers.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	The growing popularity of E-cigarettes, an alternative to cigarettes, have motivated us to study trends pertaining to the brands, ﬂavors and online market activity using posts from Reddit and Twitter. The main motivation for this world-wide study is to emphasize the difference that laws and regulations have on the usage and availability of different ﬂavors and brands of vapes in different countries. Data has been obtained from subreddits belonging to e-cigarette communities from Australia, Canada, Europe and the UK. Extensive cleaning of data, and rigorous text mining operations provide varying results for different countries. We expect the results obtained from Twitter and Reddit to be different, given the different at- mospheres they provide to users. Each tweet allows 280 char- acters for users whose identities are known, whereas Reddit posts are limited to 40,000 characters and user identities re- main anonymous. In the future, this work can be extended to posts apart from the English language.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	personal use. This difference between countries makes it im- portant to understand that the results obtained in each case need not be the same, considering the laws and regulations pertaining to e-cigarettes arent the same either.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	In America, JUUL is currently the top-selling e-cigarette brand. Other examples include the MarkTen Elite, Aspire, Kanger and the PAX Era. They are available in different ﬂa- vors including fruit, candy and creme. So far, there has been an increasing interest in ﬁnding a relation between vape ﬂa- vors and their respective health issues Li et al.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	For this project, the focus is on Canada, Australia, the United Kingdom and countries in Europe. Data collected from Reddit and Twitter provides information from vari- ous vape users writing posts about e-cigarette brands, their prices and their popularity. With this information, online markets can be found. These online markets, grouped by location, provide information on which states in countries have the highest e-cigarette online purchasing activity.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Introduction The growing importance of combating usage of e-cigarettes is evident by the increasing death tolls in the USA. A 2018 National Academy of Medicine report discovered that young people who use e-cigarettes may be more likely to smoke cigarettes in the future. As of November 21st 2019, CNBC reports vaping lung disease case total rises to 2290, while the death toll reaches 47. The reason for hospitalization has been termed as EVALI by medical professionals - E-cigarette, or Vaping, product use Associated Lung Injury. These ﬁndings make it paramount to understand the market for e-cigarettes in various countries which in turn can help the concerned authorities on taking necessary steps to curb it’s usage.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Due to the anonymity of Reddit, there are a countless number of personal vaping experiences shared by users. Hence, it is possible to study patterns and trends JP et al related to brands and ﬂavors. It would also be interesting to study the user proﬁles of e-cigarette users. However, unless explicitly mentioned, anonymity makes it hard for us to de- termine the age and gender, as both remain unknown. For this reason, speciﬁc subreddits that users subscribe to or fol- low have to be analysed to provide an estimate of what age group they belong to.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	The concept of studying alcohol and drug usage through so- cial media(Luo et al. 2016) shows how user demographics are studied to observe a trend between the time and location of Instagram posts. Additionally, the age and gender of the Instagram user is considered. Accounts followed by these users are also studied to obtain a common interest these users may share. This motivated us to perform an extensive analysis on Reddit posts and comments, to examine which ﬂavors and brands e-cigarette users might be interested in.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	A study conducted by Majmundar A et al.(2019) pro- vides us with an insight of the serious health risks associ- ated with second-hand and third-hand smoke. These insights can further be used to strengthen policies related to public e- cigarette usage and strengthen nicotine and tobacco regula- tions. They collected relevant tweets using Twitters Stream- ing API and ﬁltered the tweets by using 26 vape related terms. This study was limited by collected data from only the year of 2018. Also, the study cannot be generalized to other social media platforms, Twitter was the only platform in consideration. This motivates us to choose 2 social media platforms - Reddit and Twitter.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Sharma et al. (2016) discusses the importance of online discussions about e-cigarettes and mental health. Data is col- lected from Reddit which is freely accessible in the public domain. It was deduced that the motivations for using e- cigarettes vary between self-medication, to quit smoking, as a hobby, and to feel a sense of social connectedness.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Dai, H., & Hao, J. (2017) study also uses Twitter as a source of data by collected e-cigarette related discussions. It also emphasizes on the importance of realizing the differ- ence between human users and social bots while using social media to collect public-health related data. Opinion polari- ties motivates us to attempt a sentiment analysis, or obtain a metric of popularity of e-cigarette related laws, brands and ﬂavors by measuring favorite counts, retweets and upvotes on Twitter and Reddit.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	In a different study,data was collected from all over the world between the years 2011 and 2015(Wang L et al. 2015). Popularity of ﬂavors is measured by counting the number of times each ﬂavor was mentioned in a post. We also measure popularity in a similar manner, however, we will not con- sider the entire dataset all at once. A more thorough analysis can be performed by grouping data based country-wise.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	To collect data from Twitter, the Python library Tweepy is used to access the Twitter API. Streaming tweets with Tweepy provides a live-feed of tweets that are collected which can be ﬁltered based on keywords speciﬁed by the user. Some keywords we used to ﬁlter tweets and obtain rel- evant data are given below.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	The results are obtained in a JSON format which has been written to a .txt ﬁle. Using Pandas, an open-source data anal- ysis tool offered by Python, we can read the ﬁle into a data frame to perform further analysis.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Some major limitations of this method is the streaming limit. Twitter API allows a maximum of 3200 tweets for ex- traction per request, so the data collection takes a tremen- dous amount of time.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	The tweets obtained contain numerous special characters, user names, retweets content, URLs and emojis. The clean- ing process is extensive, making use of the Emoji and re packages offered by Python to simplify the process. Once the tweets are entirely cleaned, they are ready to be used for text mining to obtain useful information.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Data from Reddit is collected using the PushShift and Praw API. Praw is the Python API Reddit wrapper and is used to access Reddit data. Like any other text retrieval process, the data at the ﬁrst step is not entirely clean. The cleaning process includes removing special characters and stop words, lemmatization and some basic text mining procedures.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Subreddits speciﬁc to Europe, Canada, Australia and the United Kingdom were chosen in order to fully understand e- cigarette users speciﬁc to their own country. The Canadian subreddit had the highest number of users. Rest of the sub- reddit communities had approximately the same number of users.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	From ﬁgure 1, it is evident that Canada has the post num- ber of posts related to e-cigarettes. This could be attributed to more active members in the subreddit, and fewer amount of laws and regulations pertaining to e-cigarettes, thereby freely allowing e-cigarette users to explore more brands and ﬂavors and further post about their experiences online.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Data Collection As mentioned in the Data section of this paper, the data is collected from Twitter and Reddit using their respective APIs. Data from Reddit includes both the posts and their as- sociated comments. Since comments also contain signiﬁcant information, we felt like it is important to utilize it as well.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Data Cleaning The data is cleaned and manipulated per our requirements by extensively using the Pandas library on Python.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	The ﬂavors have been handpicked after reading numerous articles online. Once we shortlisted 10-15 ﬂavors indicated in Table 3, popularity of these ﬂavors were determined from Reddit posts. Here, popularity is deﬁned by the number of tweets or posts the speciﬁc ﬂavor has been mentioned in.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Brands First, we used the BeautifulSoup package offered by Python to scrape a web page by the name of misthub to get the names of all companies that manufacture e-cigarettes. Next, we determined the top 5 most common brands based on their number of occurrences on Reddit. These brands include As- pire, Kanger, Eleaf, One Up Vapor and Innokin. Surpris- ingly, Juul - despite being very popular in the USA among teenagers and young adults, was not found to be too common in other countries on Reddit. However, we have included it in the study to analyse this difference in popularity.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Online Markets The E-Cig Classiﬁeds subreddit is an online community where users buy and sell vapes. Analysing data from this subreddit helps gain a better insight of which areas have the highest online buying/selling activity when compared to oth- ers. They have a unique way to deﬁne the title of posts. Any user who buys or sells vaping products has to specify their location i.e. particular state in the US etc. This information was very helpful in determining locations with high vaping activity.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Experiments and Discussions Data Analysis of Online Markets - Reddit From Figure 2,in the USA, California, New York, Texas, Mississippi, Florida, Washington and Kentucky are some prominent states. Interestingly, according to a September 2019 article by the USNews, Kentucky is the #2 state in America having the highest e-cigarette consumption.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Figure 2: Word Cloud of States in the US based on the strength of online e-cigarette markets present.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Figure 3: Word Cloud of Provinces in Canada based on the strength of online e-cigarette markets present.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Figure 3 shows provinces in Canada having the high- est buying/selling online activity include Ontario, British Columbia, Alberta, and Quebec, to name a few.
Processed_Seeing_through_the_smoke_:_a_world-wide_comparativ.txt	Figure 4 shows that in Europe (including the UK), the United Kingdom dominates the online vape market pres- ence, followed by Germany, France, Norway, Netherlands, et cetera.

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.