Detection of 84-GHz class I methanol maser emission towards NGC 253
Tiege P. McCarthy, Simon P. Ellingsen, Shari L. Breen, Maxim A. Voronkov, Xi Chen
aa r X i v : . [ a s t r o - ph . GA ] O c t Draft version October 16, 2018
Typeset using L A TEX twocolumn style in AASTeX61
DETECTION OF 84 GHZ CLASS I METHANOL MASER EMISSION TOWARDS NGC 253
Tiege P. McCarthy,
Simon P. Ellingsen, Shari L. Breen, Maxim A. Voronkov, and Xi Chen
4, 5 School of Natural Sciences, University of Tasmania, Hobart, Tasmania 7001, Australia Australia Telescope National Facility, CSIRO, PO Box 76, Epping, NSW 1710, Australia Sydney Institute for Astronomy (SIfA), School of Physics, University of Sydney, NSW 2006, Australia Center for Astrophysics, GuangZhou University, Guangzhou 510006, China Shanghai Astronomical Observatory, Chinese Academy of Sciences, Shanghai 200030, China
ABSTRACTWe have investigated the central region of NGC 253 for the presence of 84.5 GHz (5 − → E) methanol emissionusing the Australia Telescope Compact Array. We present the second detection of 84.5 GHz class I methanol maseremission outside the Milky Way. This maser emission is offset from dynamical centre of NGC 253, in a region withpreviously detected emission from class I maser transitions (36.2 GHz 4 − → E and 44.1 GHz 7 → A + methanollines) . The emission features a narrow linewidth ( ∼
12 km s − ) with a luminosity approximately 5 orders of magnitudehigher than typical Galactic sources. We determine an integrated line intensity ratio of 1 . ± . Keywords: masers – radio lines: galaxies – galaxies: starburst – galaxies: individual (NGC253)
Corresponding author: Tiege [email protected]
McCarthy et al. INTRODUCTIONMethanol maser emission provides a powerful toolfor investigating star-formation regions within theMilky Way. Methanol maser transitions are empir-ically divided into two classes, based on the pump-ing mechanism responsible for population inversion(Batrla et al. 1987; Menten 1991). Class I methanolmasers are pumped via collisional interactions, whereasthe class II transitions are radiatively pumped. Bothclasses of methanol masers are commonly observedtowards Galactic star-formation regions, with over1200 unique sources reported (e.g. Ellingsen et al.2005; Caswell et al. 2010, 2011; Voronkov et al. 2014;Breen et al. 2015; Green et al. 2010, 2012, 2017). Theradiatively pumped class II masers are observed to beexclusively associated with young-stellar objects (YSOs)in high-mass star-formation regions (Breen et al. 2013),while class I masers are observed further from the ex-citation sources in shocked gas from molecular out-flows or expanding H ii regions (Kurtz et al. 2004;Cyganowski et al. 2009, 2012; Voronkov et al. 2010,2014).Extragalactic methanol masers have only been re-ported in a handful of sources across both classes.Class II emission has been observed (in the 6.7 and12.2 GHz transitions) towards two local group galax-ies, the Large Magellanic Cloud (LMC) and M31(Green et al. 2008; Ellingsen et al. 2010; Sjouwerman et al.2010) and the nearby starburst NGC 4945 (Ellingsenet al., submitted). There are 6 sources outside of theMilky Way with detections of class I methanol masers,NGC 253, Arp 220, NGC 4945, IC 342 and NGC 6946at 36.2 GHz (Ellingsen et al. 2014; Chen et al. 2015;McCarthy et al. 2017; Gorski et al. 2018), NGC253 at44.1 GHz (Ellingsen et al. 2017a), and NGC 1068 at84.5 GHz (Wang et al. 2014). The extragalactic class IImethanol maser emission towards both M31 and theLMC appears similar to the class II masers observed inour Galaxy, indicating they are simply bright examplesof Galactic-style class II methanol masers. In contrast,the extragalactic class I emission appears to show sig-nificant departure from the properties of their Galacticcounterparts, with the current hypothesis being theyoccur due to large-scale low velocity shocks driven bydynamical processes such as molecular in-flow alonggalactic bars (Ellingsen et al. 2014, 2017a; Gorski et al.2018). These molecular in-flows may be related to thestarburst/star-formation of the host galaxy, presentingthe possibility that the methanol maser emission scaleswith star-formation (Chen et al. 2016; Ellingsen et al.2017a; Gorski et al. 2018). The plethora of transitions and unique pumping mech-anisms between the two discrete classes of methanolmasers makes them useful tools for studying variousastrophysical phenomena. Class II methanol masershave been shown to be only associated with high-massstar formation regions and hence can be used as adirect signpost for the location of the newly formedhigh-mass star (Breen et al. 2013). They have alsobeen used to investigate the kinematics of the gas inthese regions at high resolution (e.g. Goddi et al.2011). The class I transitions trace outflows andshock fronts in the extended regions of these stellarnurseries (Kurtz et al. 2004; Cyganowski et al. 2012;Voronkov et al. 2014; McCarthy et al. 2018). The var-ious transitions of methanol can also be used to inves-tigate potential changes in the fundamental physicalconstants, such as the proton-to-electron mass ratio(Levshakov et al. 2011). Observations and comparisonof multiple methanol maser transitions from distantgalaxies will allow investigation into any variations inthis mass ratio across cosmic time (see Section 4.2 forfurther discussion).NGC 253 is a barred-spiral starburst galaxy in thenearby Sculptor group. Due to its proximity (3.4 Mpc;Dalcanton et al. 2009) and starburst nucleus, it has beenextensively studied across a broad range of wavelengths,and displays emission from numerous molecular species(Mart´ın et al. 2006; Meier et al. 2015; Ellingsen et al.2017a). A star-formation rate of ∼ . ⊙ yr − hasbeen determined for the central starburst region ofNGC 253 (Bendo et al. 2015). A variety of extragalac-tic maser species have been observed towards NGC 253,with OH, H O and NH masers having been detectedtowards the nucleus (Turner 1985; Henkel et al. 2004;Hofner et al. 2006; Ott et al. 2005; Gorski et al. 2017)along with 36.2 and 44.1 GHz methanol and HC Nmasers distributed across the central molecular region(Ellingsen et al. 2017b,a).H¨uttemeister et al. (1997) used the IRAM 30 m tele-scope to map the central region of NGC253 in a numberof methanol and formaldehyde transitions in the 3 mmband. They detected 3 different methanol transitionstowards this region, the 5 − → E (84.5 GHz line),2 k → k (96.7 GHz) and 3 k → k (145.1 GHz) lines.Although H¨uttemeister et al. found it difficult to recon-cile the properties of the 84.5 GHz line with the otherobserved thermal transitions, they did not consider thatit may be the result of maser emission (despite it beinga class I transition in Galactic star-formation regions).Given the recent detection of maser emission from therelated 36.2 GHz methanol transition, higher resolutionobservations of this 84.5 GHz transition were warranted OBSERVATIONSThe 3 mm observations were made using the Aus-tralia Telescope Compact Array (ATCA) during Di-rector’s time on 2018 July 27 and 30 (project codeC3167). The observations used a hybrid array config-uration (H75) with maximum and minimum baselinesof 89 m and 31 m respectively. Antenna 6 is not fit-ted with a 3 mm receiver, therefore, it was excludedfrom the observations. The Compact Array BroadbandBackend (CABB ; Wilson et al. 2011) was configured inCFB 64M-32k mode. This mode consists of two 2 GHzIF bands, with 32 ×
64 MHz channels, and up to 16of these 64 MHz channels can be configured as zoombands with 2048 × . ± .
001 MHz (Zuckerman et al.1972; Xu & Lovas 1997). The 31 . − at 84.5 GHz. Along with the 5 − → E methanol line,we also observed the 87.9 GHz HNCO 4 → transi-tion (rest frequency of 87 925.238 MHz; Turner 1991).In addition to the primary 3 mm observations, wealso observed NGC 253 at 7 mm on 2018 August 1during Director’s time in order to monitor emissionfrom the 36.2 GHz 4 − → E methanol maser line.The observing configuration was not optimal for high-resolution imaging, as the compact configuration (H75)does not provide high enough angular resolution at 7 mm(compared to already existing data on the 36.2 GHzline). The CABB configuration for these observationswas identical to the 3 mm observations described previ- ously, with an adopted rest frequency of 36 169 . ± .
001 MHz (Voronkov et al. 2014) and a spectral reso-lution of 0.26 km s − .PKS B1921-293 was used as the bandpass calibratorfor both sets of observations, flux-density was calibratedwith respect to Uranus (for 3 mm observations) and PKSB1934-648 (for 7 mm), while the phase calibration sourcewas PKS B0116-219. The observing strategy interleaved600 seconds on the target source with 100 seconds onthe phase calibrator. The data were corrected for atmo-spheric opacity and the absolute flux density calibrationis estimated to be accurate to better than 30%. Sys-tem temperatures were tracked using measurements ofa paddle at ambient temperature for the 3 mm observa-tions (relying on assumptions of unchanged atmosphericconditions between paddle scans), whereas a noise diodeand model for atmospheric opacity (in miriad ) was usedat 7 mm. The rms pointing error across all telescopesutilised was approximately 3.8 arcseconds. The totalon-source time for NGC 253 is 307 and 247 minutes, for3 mm (both days) and 7 mm observations respectively. miriad was used for data reduction, following stan-dard techniques for the reduction of ATCA 3 mm and7 mm spectral line observations. Self-calibration wasperformed on the data using the continuum emissionfrom the line-free channels toward the central region ofNGC 253 (both phase and amplitude). The continuumemission was subtracted from the self-calibrated uv-datawith the uvlin task, which estimates the intensity oneach baseline from the line-free spectral channels. Thisenables us to isolate any spectral line emission from con-tinuum emission. The systemic velocity of NGC 253 is243 km s − (Barycentric; Koribalski et al. 2004) and our3 mm image cube covered a range of 0 to 500 km s − .The spectral line cube for the 36.2 GHz line covers avelocity range of 120 to 470 km s − . However, thisnarrower range was still sufficient to cover the wholevelocity range over which 36.2 GHz methanol emissionhad previously been observed (Ellingsen et al. 2017a).The spectral line data was resampled and imaged witha channel width of 6 km s − for both of the frequencysetups. The miriad task imfit was use to determine po-sitions and peak flux densities of the spectral line emis-sion. This task reports the peak value and location of atwo-dimensional Gaussian fit for the emission in a givenvelocity plane within the spectral line cube. Because itreports the parameters of the Gaussian fit, minor differ-ences may be present between the reported flux-densityvalues, and those apparent from the extracted spectra.A Brigg’s visibility weighting robustness parameter of 1was used when generating the spectral line cubes with miriad . This resulted in a synthesised beam for our McCarthy et al. . ′′ × . ′′ − . ◦ , and 17 . ′′ × . ′′ − . ◦ for our 7 mm observations. RESULTSWe have detected 84.5 GHz (5 − → E) and36.2 GHz (4 − → E) methanol emission and 87.9 GHz(4 → ) HNCO emission along with 3 mm and 7 mmcontinuum emission towards NGC 253. Details of thedetected spectral line and continuum, emission are tab-ulated in Table 1 and Table 2 respectively.The 3 mm continuum emission is approximately cen-tred on the nucleus of NGC 253. Compared to the 7 mmcontinuum peak observed here, and by Ellingsen et al.(2017a), we see the 3 mm peak offset to the west by ap-proximately 0.5 arcseconds. Given that this offset is onthe order of the typical astrometric uncertainty achievedwith ATCA, we cannot state with certainty that the off-set between the 3 mm and 7 mm continuum sources isthe result of a real spatial separation.The 84.5 GHz methanol emission is located north-east of the continuum peak (see Figure 1). This loca-tion corresponds to the 36.2 GHz methanol maser regionMM4/MM5, as defined by Ellingsen et al. (2017a). The84.5 GHz methanol is not only observed from the samelocation, it also covers a very similar velocity range tothe 36.2 GHz methanol masers at this location (see Fig-ure 2). The spectrum appears to result from a singlecomponent at ∼
187 km s − which is in good agreementwith the spectral profile presented in H¨uttemeister et al.(1997), although the peak flux density obtained for theATCA data is a factor of 6 lower than that reportedby H¨uttemeister et al. (assuming a conversion factor of6.0 Jy/K for mid 90s IRAM data).Multiple components of 87.9 GHz HNCO emission areobserved, all offset from the 3 mm continuum peak emis-sion and approximately aligned with the plane of thedisk (see Figure 1). The locations of the componentsare consistent with the higher-resolution ALMA obser-vations of the same transition by Meier et al. (2015),and the lower frequency transition by Ellingsen et al.(2017a), with a component of the HNCO emission ob-served at the location of the 84.5 GHz methanol emis-sion. We have labelled the 5 components of HNCO asA through E (see Figure 1).The 36.2 GHz methanol maser transition was detectedtowards NGC 253 both north-east and south-west ofthe continuum emission at approximately the same lo-cations described in Figure 1 of Ellingsen et al. (2014).We have not included an image of this line due to thecoarse angular resolution of our 36.2 GHz observations(in comparison to existing 36.2 GHz observations to- wards NGC 253), however, the spectra at the two loca-tions are shown in Figure 3. These observations havebeen included here as they are contemporaneous withthe observations at 84.5 GHz and show that the emissionfrom the 36.2 GHz transition is largely unchanged fromprevious observations (Ellingsen et al. 2014, 2017a).Further comparison between 84.5 and 36.2 GHzemission presented in the following sections uses theintermediate resolution (original synthesised beam of3 . ′′ × . ′′ . ◦ ) data presented inEllingsen et al. (2017a) restored with the same synthe-sised beam size as the 84.5 GHz observations (6 . ′′ × . ′′ − . ◦ ). This higher resolutiondata, convolved with a restoring beam with the samedimensions as the 84.5 GHz data allows for a more ac-curate comparison of the two transitions. However, asthe two transitions were observed at different epochs,some uncertainty remains. DISCUSSION4.1.
The nature of the 3 mm methanol: masing orthermal
Class I methanol maser emission has previously beendetected in the 36.2 and 44.1 GHz methanol class I tran-sitions towards NGC 253 (Ellingsen et al. 2014, 2017a).The 84.5 GHz methanol transition we observe here isfrom the same transition family as the 36.2 GHz emis-sion. In Galactic sources, these two lines are frequentlyobserved together as masers (Breen et al. in prepa-ration). This, combined with the fact that 84.5 GHzmaser emission has been previously detected in an ex-ternal galaxy (Wang et al. 2014), the single dish detec-tion of 84 GHz methanol by H¨uttemeister et al. (1997),and our tentative detection from the Mopra 22 m tele-scope, prompted our search for 84.5 GHz emission to-wards NGC 253. Our angular resolution is a factor of 5times higher than the H¨uttemeister et al. (1997) obser-vations and allows us to better constrain the nature ofthe emission. Like the previous detections of methanolmaser emission, we need to adequately justify whetherthe emission we are observing here is due to thermal ormaser processes.The offset region from which we observe the 84.5 GHzmethanol emission is also the location of previously re-ported 36.2 and 44.1 GHz masers (MM4; Ellingsen et al.2017a). The 36.2 and 84.5 GHz methanol emissions ex-tends over the same velocity range, with the peak flux-density occurring at approximately the same velocity(see Figure 2). The 44.1 GHz maser component is alsoat approximately at the same velocity as these othertwo transitions, however, it covers a much more narrowvelocity range (Ellingsen et al. 2017a). Ellingsen et al. Table 1.
Properties of methanol and HNCO emission from our observations towards NGC 253. The locations and peak fluxdensities given are those of the peak emission components in each transition as extracted using the imfit miriad task on thespectral line cubes. Positional uncertainties (from fitting to the emission components) for the 3 mm emission is accurate toapproximately 0.5 arcsecond, and 7 mm emission to approximately 1 arcseconds. All tabulated velocities are with respect to theBarycenter.
Rest Location R.A. Dec. Peak Integrated Peak Velocity RMSFrequency Reference (J2000) (J2000)
Flux Density Flux Density Velocity Range Noise (GHz) h m s ◦ ′ ′′ (mJy) (mJy km s − ) (km s − ) (km s − ) (mJy)CH OH 84.521206 [1]
B 00 47 33.65 −
25 17 12.6 30.2 693 ±
50 187 134 – 216 2.636.169238 [3]
NE 00 47 33.90 −
25 17 11.6 30.0 961 ±
36 211 163 – 235 0.5SW 00 47 32.00 −
25 17 28.0 27.0 1372 ±
46 304 262 – 355 0.5HNCO 87.925238 [2]
A 00 47 33.94 −
25 17 11.0 53.9 953 ±
76 205 187 – 223 5.1B 00 47 33.49 −
25 17 13.8 41.3 458 ±
42 181 163 – 187 4.0C 00 47 32.79 −
25 17 21.9 23.7 276 ±
50 289 277 – 301 4.3D 00 47 32.29 −
25 17 20.5 67.5 1502 ±
88 337 319 – 361 4.6E 00 47 31.96 −
25 17 27.2 50.3 1614 ±
125 301 271 – 331 3.7Note: [1]
Xu & Lovas (1997), [2]
Turner (1991) and [3]
Voronkov et al. (2014)
Table 2.
Properties of continuum emission from our observations towards NGC 253. Positional uncertainties (from fitting to theemission components) for the 3 mm emission is accurate to approximately 0.5 arcsecond, and 7 mm emission to approximately1 arcseconds.
Frequency R.A. Dec. Flux Density RMSRange (J2000) (J2000)
Peak Integrated Noise (GHz) h m s ◦ ′ ′′ (mJy) (mJy) (mJy)3 mm continuum 84 . − .
536 00 47 33.06 −
25 17 18.4 149 217 ±
40 0.67 mm continuum 36 . − .
160 00 47 33.06 −
25 17 18.4 369 380 ±
24 0.1 (2017a) present comprehensive arguments for both whythe 36.2 and 44.1 GHz emission in these locations aremasers, and that they can not be the result of large-scaleemission from numerous Galactic-style masers. In ad-dition, high-angular resolution JVLA observations (0.1arcseconds) of this region have shown that the bright-ness temperature of some of the 36 GHz emission at thislocation, exceeds 1000 K, demonstrating conclusivelythat it is a maser (location B in Chen et al. 2018). Thesimilarity between the emission from all three of thesetransitions indicates the 84.5 GHz emission is likely theresult of a maser process.The median integrated flux density of Galactic84.5 GHz class I maser sources is 29.4 Jy km s − (Breenet al. in preparation). This corresponds to an isotropicluminosity of 735 Jy km s − kpc (5 . × − L ⊙ )at a distance of 5 kpc. Assuming a distance of3.4 Mpc to NGC 253 (Dalcanton et al. 2009), we determine an isotropic luminosity for this region of6 . × Jy km s − kpc (0 .
046 L ⊙ ), almost nine thou-sand times more luminous than the typical Galacticsource. This implies that the luminosity of 84.5 GHzemission from this single region towards NGC 253, ex-ceeds the combined luminosity of all Galactic 84.5 GHzemission.The peak emission component at 84.5 GHz has aFWHM of approximately 12 km s − . Linewidths of thethermal emission from this location are much broaderthan what we are observing from the 84.5 GHz methanolemission (H¨uttemeister et al. 1997; Leroy et al. 2015).This narrow linewidth and higher peak flux density,when compared to the 36.2 GHz maser emission, suggestthat this emission comes from a region with a smallertotal volume and higher gain. Typically maser pro-cesses are required in order to observe narrow linewidthsfrom transitions with relatively high upper-state en- McCarthy et al. −100102030 CH OH −200204060 F l u x D e n s i t y ( m J y ) HNCOA
Velocity w.r.t. Barycenter (km s −1 ) −200204060 HNCOB D e c ( J ) A B C D E −200204060
HNCOC −200204060 F l u x D e n s i t y ( m J y ) HNCOD
Velocity w.r.t. Barycenter (km s −1 ) −200204060 HNCOE
Figure 1.
Middle: Integrated 84.5 GHz methanol emission (red contours 10%, 30%, 50%, 70%, and 90% of the 573mJy km s − beam − peak), integrated 87.9 GHz HNCO emission (black contours 15%, 30%, 50%, 70%, and 90% of the1.9 Jy km s − beam − ) and the 3mm continuum emission (white contours 2%, 10%, 30%, 50%, 70%, and 90% of the 149mJy beam − peak) with background image of integrated CO J = 2 → − ). The vertical dashed line indicates the systemic velocity of NGC 253(243 km s − ; Koribalski et al. 2004). −1 )−10−5051015202530 F l u x D e n s i t y ( m J y ) Figure 2. . ′′ × . ′′
150 200 250 300 350 400 450Velocity w.r.t. Barycenter (km s −1 )−10−5051015202530 F l u x D e n s i t y ( m J y ) Figure 3. − . Emission towards the north-eastand south-west regions are represented by the black and redspectra respectively. − → E transition consid-ered here). In addition comparison of the integratedline ratios between the 84.5 GHz methanol emission,the two other class I transitions, and 48.4 GHz ther-mal methanol (Ellingsen et al. in preparation) fromthe same region are inconsistent with a thermal process.H¨uttemeister et al. (1997) found inconsistencies in theircalculations of environmental properties when consider-ing emission from the 5 − → E line as thermal, thiswould be expected if some component of the emissionwas resulting from maser processes. From this point on,we will refer to this emission as a class I methanol maser.4.2.
Ellingsen et al. (2017a) outline the potential useof extragalactic class I methanol masers as probesof the proton-to-electron mass ratio. The rest fre-quencies of methanol transitions are sensitive to theproton-to-electron mass ratio due to the hindered in-ternal rotation of the OH radical in the methanolmolecule (Levshakov et al. 2011). Variations in thesefundamental constants can be revealed through com-parison of rest-frequencies observed astronomicallyand those observed in the laboratory (Kanekar 2011;Bagdonaite et al. 2013). For this to be practical on cos-mological scales, multiple methanol transitions need tobe identified that are sufficiently luminous, co-spatialand have different dependencies on these physical con-stants (see Section 4.4 in Ellingsen et al. 2017a, formore detailed discussion). Ellingsen et al. concludethat the 44.1 GHz methanol maser lines observed to-wards NGC 253 are not appropriate for this kind ofcomparison, due to the significant difference in luminos-ity and spectral profile when compared to the 36.2 GHzemission. From our observations the 84.5 GHz transi-tion broadly appears to be more appropriate for thissort of comparison (with sensitivity coefficients of − . − . Comparison with other class I emission
When considering the Ellingsen et al. (2017a) inter-mediate resolution observations of 36.2 GHz towards thesame region as the 84.5 GHz maser, we see two 36.2 GHzpeaks either side of the 84.5 GHz peak velocity. This isdue to the 36.2 GHz emission from this region consist-ing of two distinct components, one on the south sideand the other on the west side of the MM4 region. Atthe resolution of our observations we are not able to de-termine whether the 84.5 GHz emission is broken intotwo discrete components also, with the spectrum we areseeing representing contributions from the two compo-nents. Figure 2 shows a composite spectrum with emis-sion from both the 36.2 and 84.5 GHz transition. Thisspectrum has been taken at the location of the 84.5 GHzpeak (see Table 1), using both our 84.5 GHz data, andthe intermediate resolution 36.2 GHz spectral line cubefrom Ellingsen et al. (2017a) restored with the same syn-thesised beam size (see Section 3 for further details).The integrated flux-density of this re-convolved36.2 GHz emission is 798 mJy km s − , resulting in aintegrated intensity ratio between the 36.2/84.5 GHzlines of 1 . ± .
4. This ratio is similar to the median36.2/84.5 GHz ratio observed towards Galactic sourcesof 1.4 (Breen et al. in preparation). In order to deter-mine the 36.2/84.5 GHz line ratio across the remainingmethanol maser regions, we determine a generous up-per limit for the integrated flux density of undetected84.5 GHz emission of 326 mJy km s − (for emission with5 σ peak and FWHM of 10 km s − ). We compare thisupper limit 84.5 GHz integrated flux against 36.2 GHzintegrated flux densities from our re-convolved cube atthe locations of the 36.2 GHz components (see Table2 of Ellingsen et al. 2017a). This results in lower limitline ratios of 3.5, 4.2, 2.9, 1.7 and 1.5 for regions MM1,MM2, MM3, MM6 and MM7 respectively. This indi-cates that in all of these regions, the intensity of any84.5 GHz emission is lower relative to the 36.2 GHzemission when compared to MM4.Maser models show that both the 36.2 and 84.5 GHzmaser transitions respond similarly to the conditions oftheir environments (McEwen et al. 2014; Leurini et al.2016). Figure 2 of McEwen et al. (2014) shows thatacross the range of possible environmental densitiesfor these transitions, at the lowest viable densities the36.2/84.5 line ratio is at a maximum, with this ratiodropping and tending towards zero as density increases.This suggests that in the region we are seeing both of McCarthy et al. these transitions, the density is likely higher than in theregions of methanol we do not detected the 84.5 GHztransition. This is further supported by the fact that therange of viable densities for the 44.1 GHz maser tran-sition are shifted towards lower values (McEwen et al.2014; Leurini et al. 2016). Therefore, in a higher den-sity environment we would expect to see higher gainfrom both the 36.2 and 84.5 GHz transitions (comparedto the 44.1 GHz line), which is what we observe towardsthe 84.5 GHz maser region. From McEwen et al. (2014),the line ratio of 1 . ± . of approximately 10 –10 cm − ,whereas the upper limits of between 1.5 and 4.2 we cal-culated for the remaining regions correspond to densitiesof ∼ –10 cm − (Figure 2 of McEwen et al. 2014).Maser beaming can significantly affect line ratios be-tween observed methanol maser transitions, especiallywhen considering methanol lines from different tran-sition series (Sobolev et al. 2007; Sobolev & Parfenov2018). However, as the 36.2 and 84.5 GHz transi-tions are from the same series (first MMI regime fromSobolev et al. 2007), we expect beaming may have lessof an affect on this ratio. Ellingsen et al. (2017a) sug-gest that the class I methanol masers are predominatelyfrom low-gain, relatively diffuse emission regions. So wewould expect these masers to be at significantly lowerdensities than those characterising cold dense Galac-tic cores. However, such densities could be achievedwithin dense giant molecular clouds similar to Sagitar-ius B2 (Huttemeister et al. 1993). It should be notedthat the models from both McEwen et al. (2014) andLeurini et al. (2016) focused on investigation of Galac-tic environments (primarily high-mass star formation re-gions), and it may be possible there is a low densityregime where the 36.2/84.5 GHz line ratio approachesunity which would be more consistent with diffuse maseremission. 4.4. Future prospects